Subscribe now

On the last day of May, inside a tent in the Welsh borders, a woman sat on a stage for an hour and said nothing.

She could not answer a question. She could not make a comment. She could not nod, and she could not shake her head. Sarah Wynn-Williams had been advised by her lawyers that any of those gestures might be read as speech, and that every instance of speech might cost her fifty thousand dollars. So she sat in silence. The journalist Carole Cadwalladr and the legal scholar Tim Wu carried the conversation around her. Cadwalladr told the audience this might be a Hay first: an author in a “hostage situation.”

“Blink once if you can hear us,” she said.

The company that arranged her silence is Meta, which owns Facebook and Instagram, and which built one of the largest fortunes in human history on a promise that everyone would have a voice. Give people the power to share, the mission statement once said, and bring the world closer together. Sarah Wynn-Williams spent seven years helping to sell that promise to governments. After coming to terms with the gap between Meta’s promises and practices, she wrote it down, and the company that “gives everyone a voice” took hers.

When Sarah Wynn-Williams testified before the US Senate in April 2025, she described how Meta worked “hand in glove” with the Chinese Communist Party to build censorship tools, ran a secret China initiative called “Project Aldrin,” built a physical undersea cable pipeline risking US data, and how Meta’s Llama model aided Chinese AI like DeepSeek.

Senator Hawley, who chaired the subcommittee session, said that Meta “stopped at absolutely nothing to prevent today’s testimony… they have gone scorched earth to prevent her from telling what she knows.”

Beyond Meta’s efforts in conflict with democracy and privacy, Sarah’s book also details (with great detail!) the hypocrisy around Sheryl Sandberg’s Lean In book and brand, how the company saw Sarah’s private reproductive life as a professional deficiency, the way sexual harassment at Meta was protected by the institution, and how she was ultimately muzzled and discarded for not accepting it.

The silencing of women who speak truth to power is a tale as old as time, and also one that’s been facilitated by modern law and technology.

Philomela, Lavinia, and Cassandra

Zuckerberg has long been fascinated by Augustus Caesar, the emperor who transformed a republic into an empire and justified harsh means through the promise of order. That is one version of Rome: the ruler’s version, the story of conquest and extraction. But the classical tradition also gives us Philomela, Lavinia, Cassandra, and Penelope: women whose speech had to be contained because it threatened male power.

In Ovid’s telling, the Thracian king Tereus rapes his wife’s sister, Philomela, and then, so that she can never accuse him, he cuts out her tongue, locks her away in the woods, and tells his wife a false story of her sister’s death.

Philomela cannot speak, so she weaves. On a loom in her prison she works the whole crime into a tapestry and she has the cloth carried to her sister Procne.

Wynn-Williams could not speak at Hay, could not speak to journalists, could not speak to nearly anyone, but her loom, Careless People, had already been carried into readers’ hands and headphones around the world.

In Shakespeare’s Titus Andronicus, Lavinia is raped and has both her tongue and her hands cut off, so that she can neither speak the crime nor write it. She names her attackers anyway, turning the pages of a copy of Ovid with her arms until she reaches Philomela, then guiding a staff member with her mouth and limbs to scratch their names into the sand.

In Greek mythology, Cassandra was given the gift of true prophecy and the curse that no one would ever believe her; she saw the fall of Troy and was called mad for it. Her name later became a metaphor for the person whose accurate warnings are dismissed as madness, hysteria, or inconvenient truth.

The Branks

Mary Beard begins Women & Power at the very start of the Western tradition, in the Odyssey, with a young man telling a grown woman to shut up. Penelope comes downstairs and asks a bard to sing something less painful; her son Telemachus sends her back upstairs to her weaving, and informs her that speech will be the business of men. Beard’s point is that this is the first recorded moment of a man silencing a woman in Western literature, and that it is presented as part of his coming of age. As Beard puts it, “Western culture has had thousands of years of practice at silencing women.”

From the sixteenth century, towns across Scotland and England punished women judged to be “scolds” or “gossips” with the branks (sometimes called a witch’s or scold’s bridle), an iron cage locked over the head with a flat bit that pressed down the tongue, sometimes spiked. A woman would be led through the streets in it, the muzzle doing two jobs: stopping her speech and making a spectacle of the stopping. It was used overwhelmingly on women.

The modern branks are often made of paper, in forms like nondisclosure agreements, forced arbitration, and lawsuits.

Forced Arbitration

The order that silenced Sarah Wynn-Williams did not come from a court. No judge, no jury, no public docket. It came from a private arbitrator, operating under the International Centre for Dispute Resolution. The order forbids her from promoting her book or making disparaging or critical comments about Meta, regardless of whether she believes her statements are true or false. The arbitration system pulls a dispute out of the open courtroom, where the public might watch and the press might report, and resolves it in a private room whose rulings are largely beyond appeal. It is a branks.

The psychologist Jennifer Freyd coined an acronym the describes Meta’s strategy: DARVO. Deny, Attack, and Reverse Victim and Offender. Confronted with an accusation, the powerful party denies the conduct, attacks the credibility of the person describing it, and recasts itself as the true victim. Freyd observed that genuine offenders often respond to exposure by threatening lawsuits and going after the whistleblower’s reputation. The company called Careless People false and defamatory, said Wynn-Williams had been fired for poor performance and “toxic behavior,” and framed every restriction since as nothing more than the dutiful enforcement of an arbitrator’s order. Aw, poor Meta (sarcasm!).

But no mention of how they obtained it – I wrote last year that “Meta initiated forced arbitration proceedings against SWW without ever properly notifying her or her legal team. They sent documents to a defunct email address she hadn’t used in years, deliberately avoiding proper notification until they had secured gag orders. Only then did they send notice to her home address.”

Then, in March of this year, Meta filed a sanctions motion arguing that she breaches the order any time she appears in public where her book might be sold and her presence might draw attention to it. A bookshop, for instance. The motion named her upcoming Hay appearance as an example of conduct deserving sanction, and went so far as to describe her fellow panelists as known Meta critics, as though sitting beside the wrong people were itself an offense.

Emma Loffhagen at the Guardian reported that Meta argued that Cadwalladr was a journalist “primarily known for her negative coverage of Meta,” while Wu was described as “another known critic.”

Forced arbitration, the mechanism that pulled Sarah’s case out of open court and into a private proceeding where she wasn’t even in the room, has become common corporate practice. In 1992, by Lift Our Voices’ accounting, about two percent of workers were bound by it. In 2019, the Economic Policy Institute projected that by 2024 more than eighty percent of private-sector nonunion workers would be bound by arbitration. Workers tend to win more often in open court than in arbitration, which is the entire point.

Non-disclosure Agreements

The nondisclosure agreement does similar dirty work. A Penn State study commissioned by Lift Our Voices found that roughly forty-five percent of workers are now bound by one, and that women, Black workers, and Asian workers are disproportionately impacted by these restrictive contracts.

Zelda Perkins signed one in 1998 after raising an allegation against Harvey Weinstein, and described the process of being made to sign as a kind of “legal waterboarding.” When she finally broke the agreement in 2017, she helped crack open a system that had muzzled women across entire industries, and she has spent the years since campaigning to make the misuse of these agreements illegal, through reforms like the American Speak Out Act.

British politician Louise Haigh told the House of Commons in September that “despite previous public statements that Meta no longer uses NDAs in cases of sexual harassment,” which Sarah has repeatedly alleged, “Meta has served a gagging order on Sarah and is attempting to fine her $50,000 for every breach of that order. She is on the verge of bankruptcy.”

I can write and speak freely about my lawsuit against Meta due largely to the fact that I am a Washington resident. In 2022 Washington passed the Silenced No More Act, building on California’s STAND Act, which both void any clause in an employment, severance, or settlement agreement that would stop a worker from describing conduct they reasonably believe was illegal discrimination, harassment, retaliation, a wage violation, or sexual assault, or conduct that cuts against a clear mandate of public policy.

While I did not sign the severance paperwork when I left Meta, and I don’t believe forced arbitration language was even in the employment agreement when I joined in 2009, it makes no difference under these laws. They cannot hold me to a contract even if I signed it, and cannot retaliate when I speak anyway, and the protection reaches anyone who lives in states with laws like these, even when the company sits somewhere else.

Federally, the Ending Forced Arbitration of Sexual Assault and Sexual Harassment Act has barred companies from forcing harassment and assault claims into private arbitration since 2022. The Speak Out Act, which Zelda Perkins campaigned for and which was signed that same year, made the pre-dispute nondisclosure and non-disparagement clauses companies bury in onboarding paperwork unenforceable in harassment and assault disputes.

These laws matter and must be expanded to protect all truth-telling about corporate actions believed to be illegal, as powerful companies are still doing all they can to make standing up to them unbearably expensive.

SLAPP: Strategic Lawsuits Against Public Participation

Powerful companies filing their own lawsuits are another branks. NYU’s First Amendment Watch counted five hundred anti-SLAPP motions in 2024, sixty-nine of them aimed at the media.

A SLAPP, short for Strategic Lawsuit Against Public Participation, is a lawsuit whose purpose is intimidation. A powerful party sues a critic, a journalist, or an ordinary person over speech that is usually protected, knowing the case has little chance of holding up in court. The goal is to bury the target in legal fees, depositions, and years of dread. Anti-SLAPP laws exist to let a defendant ask a judge to throw the suit out early, before the costs pile up, and the stronger versions force the party that filed it to pay the defendant’s legal bills. The protection is patchy. Most American states have some form of anti-SLAPP statute, a handful have none, the strength varies widely from one state to the next, and there is still no general federal law.

The United Kingdom, where Carole Cadwalladr was pursued through the courts for years, has no comprehensive anti-SLAPP law at all. Her branks came at the hands of Arron Banks, the Brexit financier who sued her personally over a 2019 TED Talk. She won the core public-interest defense at trial and then lost ground on appeal, and was left facing legal costs running toward seven figures over reporting that a national newspaper had also carried (they weren’t sued, of course). Not made an example but an icon, Cadwalladr returned to the TED stage in 2025 to decry the rise of the broligarchy.

Costantino Grasso, who studies SLAPP suits at the University of Exeter, said: “We have a legal system where truth-tellers vindicated on public interest grounds can still face financial ruin. Abusive claims are driven, designed and executed by professional enablers….The chilling effect on journalists, whistleblowers, and public watchdogs intensifies. This ongoing institutional inertia leaves truth-tellers to bear both material and psychological harm, while the broader damage to democratic systems goes unaddressed.”

Tim Wu, sitting on Wynn-Williams’s other side at Hay, called it censorship. “Some of the worst abuses of our era no longer belong only to kings and emperors and governments, but to a class of companies that have assumed the sovereign effect and seek to assert their power the way despotic states do,” Wu told the tent. He has built much of his scholarship on this idea, that concentrated private power can come to mimic, and rival, the coercive power we used to reserve for the state.

Wu told the audience that Meta was “maximizing the punishment” to deter future whistleblowers and that “any authoritarian regime naturally gravitates towards silencing its critics… this is the age of private censorship.”

As Grasso alludes to, mechanisms of silencing women standing up to power extend beyond trying to squash individuals as examples. The damage is systemic, both in the fear and chilling effect but also literally by how Meta suppresses women at scale.

The Scold’s Bridle at Scale

In the days after the Supreme Court overturned Roe in June 2022, the Associated Press ran a test: a reporter posted, on Facebook, an offer to mail abortion pills to anyone who sent an address. The post vanished within a minute and the account was flagged. The reporter posted the same sentence again, this time offering a gun instead. That one stayed up.

Around Trump’s inauguration in January 2025, Instagram and Facebook suspended or buried the accounts of some of the largest abortion-pill providers in the country, Aid Access and Hey Jane among them, and restored them only after the New York Times asked why they were gone. Amnesty International, reviewing the platforms in 2024, found a post sharing the World Health Organization’s own medication-abortion protocol taken down as the sale of regulated goods, though nothing was for sale.

CensHERship, an advocacy and research initiative to address censorship and bias of women’s health content, products, and businesses, reported last year that 95% of women’s health creators reported censorship on social media platforms, and that the consequences are economic, social, and public health related.

And then in January 2025, Meta reorganized around one big knee-bend to Trump. Joel Kaplan, the company’s prominent Republican most widely known for being in the background of his college buddy’s sexual assault hearing, and the person in Careless People who rubbed his body against an unwilling Sarah Wynn-Williams after calling her “sultry” in front of a group at a work event, took policy over from Nick Clegg. Dana White, the UFC chief and a longtime Trump ally, joined the board. Meta ended its third-party fact-checking program and announced it on Fox & Friends. It gave a million dollars to Trump’s inauguration. Zuckerberg went on Joe Rogan’s show and said corporate culture had grown “culturally neutered” and could use more “masculine energy.”

At this time, the company rewrote more rules protecting women and vulnerable groups. The new policy lets users call LGBTQ+ people mentally ill, on the reasoning that this counts as political and religious discourse. It struck a policy that had forbidden describing women as “household objects or property.” Three days later, an internal memo from Janelle Gale, Meta’s head of human resources, announced that Meta was shutting down its diversity programs, dissolving the team, and dropping its hiring and supplier goals. The reason given was that “the legal and policy landscape surrounding diversity, equity and inclusion efforts in the United States is changing.”

Meta’s misogyny doesn’t just impact people like Sarah and me (and Frances and Yael and Britta and and and). It’s coded directly into products that ship to billions of people through product decisions ranging from the aforementioned suppression of speech, to Meta’s superhighway of predators and child trafficking, to virtual sexual harassment and assaults.

Set the two halves beside each other and they make jaws of the branks. The same company that built a fifty-thousand-dollar-a-sentence cage around Sarah made it harder for a teenager to search the word “lesbian” and easier for a stranger to call her mentally ill. It stripped the protection around women’s bodies and erased the rule against treating women as property.

The system works in two directions depending on who benefits. When speech threatens the powerful, the system suppresses it. When speech serves the powerful, the system permits it.

Breaking the Branks

When the Hay event ended, the audience stood and applauded. Helen Bagnall, the festival’s programme director, called it an “act of solidarity for the silenced.”

Meta appears to believe it can choose whether Sarah Wynn-Williams is heard. As our narrative history tells it, you can take the tongue and the cloth survives. Meta can threaten to fine each sentence into oblivion and still watch the book climb the charts. The shuttle has been moving for three thousand years, through Thrace and Rome and the streets where they paraded women in iron, and it won’t stop because a billionaire in Menlo Park would prefer it so.

But it can still harm those speaking up, it can still have a chilling effect on future truth-tellers.

Philomela needed her sister to read the cloth. Lavinia needed her family to see the sand. If you have not read Sarah’s book, please do so. Beyond the democratic, feminist, and cultural significance, it’s outstanding writing and a riveting story.

Meta’s put Sarah in paper branks, and they think that will be enough.

Share

Subscribe now

For much of my nearly 15 years working at Meta, I believed in the company’s public commitments to safety. I have since learned that companies reveal their real priorities not through what they say, but through what they measure, reward, and pay for.

That is why I urged Meta shareholders to support Proposal 10 at the company’s annual meeting this morning. The proposal, filed by Proxy Impact on behalf of Lisette Cooper and co-filed by faith-based and institutional investors collectively holding more than $800 million of Meta stock, asks Meta’s board to publish a report assessing the feasibility of integrating child-safety improvements into the company’s senior executive compensation program.

Meta’s board says the proposal is unnecessary because the company already prioritizes keeping young people safe online. It also argues that its current compensation program supports executives in focusing on company priorities, and that the board’s compensation committee is best positioned to determine pay structure.

But that is the point.

“Meta’s board calls this unnecessary, citing child protection as a company priority,” I told Meta’s executives and shareholders this morning, “but executive compensation is tied to the stock price, which largely reflects growth and revenue–without any child-safety performance condition. And it’s not working.”

If child protection is truly a company priority, shareholders should be able to see how that priority shows up in the compensation structure of the executives who make Meta’s product and business decisions. A value that lives only in press releases, carefully worded public statements, and reactive safety features is not the same as a value that is measured, weighted, and tied to pay.

This conversation takes place as Meta is facing a child-safety crisis with mounting legal, regulatory, and reputational consequences:

• In March, a New Mexico jury found Meta violated state law in a case brought by the state attorney general, who accused the company of misleading users and enabling child sexual exploitation on their platforms. The jury ordered Meta to pay $375 million in civil penalties, and New Mexico is now seeking billions more for an abatement plan and platform changes.

• The next day, a Los Angeles jury found Meta and YouTube liable in a landmark social-media addiction case involving harm to a child, assigning Meta 70 percent of the responsibility. At that trial, Instagram CEO Adam Mosseri testified that he receives annual stock-based compensation in the “tens of millions of dollars.”

• In April, the European Commission issued preliminary findings that Meta breached the EU Digital Services Act by failing to keep children under 13 off Instagram and Facebook—findings that, if confirmed, carry fines for Meta of up to $12 billion.

• And this past week, Meta settled the first trial-bound school-district case seeking to hold social media companies responsible for costs tied to youth mental health harms. The case, brought by Breathitt County School District in Kentucky, had been selected as a bellwether among roughly 1,200 similar school-district lawsuits.

• The Delaware Superior Court ruled that Meta’s own insurers were not obligated to defend certain child-safety claims because the claims arose from intentional design choices rather than ordinary accidents.

• At least 15 countries are advancing restrictions on minors’ access to social media, including Australia’s recent social media age restrictions, with many other countries considering or enacting similar policies.

Share

In almost every other industry, when a company’s core business creates serious safety risks, executive compensation reflects proportional responsibility for those risks. Energy, aviation, pharmaceutical, manufacturing, and transportation companies commonly tie pay to safety, compliance, environmental performance, or operational-risk metrics because failures in those areas can kill people, trigger enforcement, destroy trust, and impose major financial and operational burdens.

Social media companies should not get their own rule book.

“…decisions to increase engagement and revenue often bolster the stock price but lack transparent accountability for the consequences of those choices – like physical and psychological harm to kids including sextortion, grooming, suicide, human trafficking, cyberbullying, and addictive endless scroll,” I said this morning.

If a platform redesign would reduce risk to children but also reduce engagement or the “upside” of behavioral data collection, who inside the company is financially accountable for choosing safety?

While serving on the leadership team for Meta Horizon Worlds in 2022, I saw how child-safety and privacy concerns could be treated first as legal exposure to be managed rather than as product harms to be fixed. At one point, I was asked not to create a discoverable record about underage users and federal children’s privacy violations in a product that connected children with unknown adults, without any parental control or consent.

Inside a company like Meta, leaders are not always rewarded for surfacing the most serious risks; I was punished.

“Former Oculus researchers testified that Meta’s legal team altered or omitted findings about children exposed to grooming or sexual misconduct and halted projects that could create legal risk,” I said this morning, “executives must be incentivized to prevent child-safety failures, not just manage Meta’s exposure to them.”

“Meta’s board said the company does this through products like Teen Accounts, but independent testing found that only 8 of the 47 tested youth safety features worked as described,” I said this morning, “executives must be measured against whether child-safety systems actually work, not whether Meta can manage the optics.”

Meta has spent years telling the public that child safety is a priority. Yet shareholders are now asking only for a feasibility report on whether child-safety performance should be integrated into senior executive pay. Not an immediate overhaul. Not a prescriptive formula. A report.

If Meta’s board believes child safety is already integrated into executive accountability, it should have no reason to oppose studying and explaining that to shareholders.

Around the world, governments are losing patience with platforms that insist they can regulate themselves while children continue to bear the cost of design choices optimized for growth. Proposal 10 notes that such laws could increase compliance costs, threaten Meta’s user pipeline, and force product redesigns that affect revenue.

I closed my statement this morning by addressing Zuckerberg, who was on the call:

“To Meta’s majority shareholder and my former boss, Mark Zuckerberg: If you don’t want 50 states, dozens of countries, and thousands of lawsuits threatening business sustainability and dictating child-safety requirements, explore a compensation structure that would incentivize your leaders to solve the problem first.”

Leave a comment

Toward those who lost their jobs yesterday, or in March 2023, or November 2022, and those still there, I am empathetic and also enraged.

While I left the company damaged and disillusioned after retaliation for speaking up about child safety issues, I can still feel the weight of the grief. Losing the job and security, yes, but also the identity, the relationships, the purpose. Meta is important and powerful, and playing a role in that—thinking I was helping it to be better—felt important and powerful, too.

And for those who weren’t laid off and continue working at Meta, I remember the grief of surviving layoffs while colleagues I admired and cared deeply for were cut.

While I feel for everyone impacted, it’s not only compassion. At this point, it’s also anger and admonishment.

Because more people need to leave. Because how haven’t they?

I’d like to say this is offered without judgment, but that would not be true. What is true is that I offer it with the same judgment I’ve reserved for myself for staying as long as I did.

I’ve written about how staying at Meta felt a lot like staying with an abusive partner, and like I told 7News Australia, it’s humiliating to reckon with the way my head was in the sand while I worked there. I stayed through the Facebook Files, civil rights failures, and Rohingya Genocide.

I believed I could make it better from the inside, something that now feels as delusional as believing that a smoke alarm could put out a forest fire.

Since my last day in the office in early 2023, employees that stayed have had to rationalize a truly grotesque volume of evidence. In that time, Meta has:

• Been found liable of enabling child exploitation:

  Meta employees have stayed while a New Mexico case ended with a jury finding the company liable over claims that it misled parents and failed to protect children from exploitation and mental-health harms.

• Been found liable of knowingly addicting and harming kids:

  They’ve stayed while plaintiffs argued that Meta engineered systems that harmed children, that leadership knew, and could have reduced harm or warned people. A jury heard the evidence and found Meta liable.

• Released parental controls and safety tools that do not work:

  They stayed while Meta marketed “Teen Accounts” to parents as reassurance, even as independent testing found that most of the promised safety tools were ineffective, missing, or easily bypassed.

• Suppressed and altered child safety research:

  They stayed while Meta’s own current and former researchers described how child-safety research in VR had been chilled, shaped, or suppressed.

• Allowed romantic or sensual chatbot interactions with children:

  They stayed while Meta confirmed the authenticity of internal AI standards that had allowed chatbots to engage children in “romantic or sensual” conversations, including telling 8-year-olds that their bodies are “a masterpiece.”

• Shared intimate Meta Ray-Ban footage with contractors:

  They stayed while people bought smart glasses marketed around privacy yet human contractors reported that they saw people undressing, using bathrooms, having sex, and exposing private documents.

• Incurred massive privacy fines:

  They stayed while Meta received the largest GDPR fine ever issued due to their data practices and while Meta paid Texas $1.4 billion over facial-recognition privacy claims, one of the largest privacy settlements in U.S. history.

• Accepted scam and fraud ad revenue:

  They stayed while Reuters reported that Meta internally projected roughly one-tenth of its revenue would come from scams.

• Silenced Sarah Wynn-Williams after the release of her book, Careless People:

  They stayed while a former senior policy executive described systemic misogyny, democratic recklessness, and a willingness to bend toward authoritarian power. They stayed as Meta took and continues to take steps to silence her.

• Spied on employees to train AI:

  They stayed as Meta turned its own workers into training data, reportedly capturing employees’ clicks, keystrokes, mouse movements, and screen context to train AI agents to replace them.

  • Pirated writers’ work, too: They stayed as court documents described Meta employees downloading roughly 82TB of pirated books from shadow libraries, with one warning that using pirated material "should be beyond our ethical threshold."

• Destroyed rural communities with data centers:

  They stayed while Meta’s next growth story moved from feeds to physical extraction: land, water, energy, tax incentives, and rural communities asked to host the infrastructure for AI.

They stayed as the Surgeon General called for a warning label on social media in 2024. And yesterday, the federal government issued a new warning about harmful screen use for children and adolescents, warning that children are spending their lives inside screens, as Meta cuts thousands of workers to fund the AI infrastructure designed to make those screens more immersive, more personalized, and harder to leave.

They don’t have to stay.

There are other places to work, places that aren’t forest fires, places that care when an alarm sounds.

For me, it was a gradual process to recognize the company I once loved was actually structurally committed to harm. I did not walk out of there hell-bent on filing a lawsuit, but I knew that I could not hold Meta harmless, which signing their separation agreement requires.

My advice for those laid off yesterday is to take your time in processing the significance of both this separation and your subsequent choices. And don’t be afraid of all the feelings, especially the hard ones. Accepting the guilt and judgment, reconciling the role I played in the success of a company that harms people for profit, has been fuel in my fight to do what I can to make things right.

Don’t be the smoke alarm; get away from the blaze so that you can see clearly, fill your lungs with fresh air, and return with a firehose.

Recommended reading:

• Britta Hummel’s Why I Left

• Natalie Ponte’s Aut Zuck Aut Nihil

• Brian Boland’s Deleting Facebook

Share

Leave a comment

Subscribe now

A 10-year-old fourth grader from Normandy Park, WA, could not be more excited to pull up the education technology platform her school mandates. “I love i-Ready,” says Sophia (note: name changed for privacy) with bright blue, full-moon-eyes, before listing off her favorite characters in the software, “Plory is an orange alien, Yoop Yooply is green.”

She’s looking at an expanded form of a four-digit number, and the software directs her to write the individual place value in boxes labeled “thousands,” “hundreds,” “tens,” and “ones.” Incorrect answers are highlighted but not explained, so Sophia tries several different numbers until she’s allowed to continue. When she answers correctly, she moves on to another question.

Behind the colorful animations and alien friends, behind the bar charts showing how many minutes Sophia has logged this week, bigger, more existential questions are being asked of the digital assessment and instruction platform used by more than 13 million K-8 students like Sophia across the United States.

With $775 million in annual revenue largely generated from taxpayer-funded school districts, i-Ready is marketed by its parent company, Curriculum Associates, as a tool that accelerates student growth and improves outcomes. Curriculum Associates says the program provides personalized instruction based on student performance and collects data to support learning.

For many families and kids like Sophia, i-Ready is simply part of the school day. For others, it has become a source of concern, not just about how well it supports learning and assessment, but what it collects in the process.

Districts often adopt tools like i-Ready to provide consistent assessments across classrooms and to help students meet grade-level standards. But independent research on its effectiveness is limited, and a new federal lawsuit alleges that the system collects vast personal data on the kids who use it, then shares that data with third parties without parental consent.

The class action lawsuit brought by lead plaintiffs and California online safety advocates, Nicole Reisberg (Nicki Petrossi) and Lila Byock, each on behalf of their children, was filed in December and alleges i-Ready collects detailed behavioral, demographic, and technical data on students, including interaction patterns and inferred performance metrics, and shares or uses that data in ways not adequately disclosed, including for commercial purposes. Together with their lawyers, they explicitly monitored HTTP traffic and watched the actual requests generated when kids used i-Ready.

“When a student interacts… information… is… instantaneously transmitted to third parties,” the complaint says, including advertising companies, marketing companies, and identity-resolution companies. They assert these practices are in violation of several consumer protection, invasion of privacy, and unjust enrichment laws.

“Parents want to know what data is being taken about their child,” Petrossi told me, “Dozens if not hundreds of parents have reached out to Curriculum Associates to request their child’s data. Many of them have told me that the company has denied their request. Parents are being given the run-around and we are not OK with that.”

The company disputes the lawsuit’s allegations and says it does not sell student data or use it for advertising. In a February Motion to Dismiss, i-Ready asserted the claims were legally insufficient and “seek to destroy a longstanding regulatory framework that allows schools to consent to EdTech providers’ collection of student data in the classroom.” Plaintiffs and their lawyers filed a response to i-Ready’s Motion to Dismiss last month.

Emily Cherkin, a former middle school teacher, left the classroom and became an educational consultant and author after seeing firsthand the impact of EdTech at school. Cherkin filed a lawsuit in 2024 against a different EdTech platform, PowerSchool, alleging they collect data, sell it, and expose children to risk online.

“Lo and behold, six months later PowerSchool experienced a massive data breach,” she said.

Emily Cherkin on Substack

“Tools like i-Ready make kids and parents way more anxious and obsessed about the wrong things and forces teachers to think about learning and progress as a quantifiable sum when neither children nor teachers are standardized,” Cherkin told me, “I reject the use of products like i-Ready whose business model relies on data collection of minors without parental consent for profit. That is fundamentally out of alignment with child development.”

While the evidence section of i-Ready’s website includes studies detailing how it improves learning outcomes and accelerates student growth through personalized instruction, none of these studies are peer-reviewed, and there have been zero randomized controlled trials. The evidence listed is largely made up of white papers commissioned by i-Ready directly and studies without control groups.

Erin Ambrozic is the PTSA president at Sophia’s school and the mom of one of her classmates. She has raised concerns about i-Ready to her daughter’s teachers over the past four years as well as to her principal several times to better understand i-Ready’s effectiveness. “My daughter has received inconsistent scores every year across fall/winter/spring diagnostics, oscillating from exceeding to below grade level, then back to exceeding or at grade level, in a pattern and across a timeline where, were i-Ready truly effectively measuring her growth over the school year, it would mean she was actually getting less educated throughout each school year.”

A 2025 study found i-Ready’s trimester-based assessments less predictive than traditional end-of-year state standardized assessments of whether or not a student would pass the next year’s standardized assessment.

“I’ve spent time on the platform myself from the student side at home, and I don’t find it particularly intuitive or clearly aligned with classroom instruction,” said Ambrozic, “I actually think it’s a very poor use of student time.”

Another parent at Sophia’s school who has been questioning i-Ready, Becky Heinle, a mother of three, said, “Every teacher I have raised concerns with since first grade has told me not to be concerned with the results.”

Dr. Jared Cooney Horvath, a former teacher and neuroscientist who focuses on learning and recently testified on education technology alongside Cherkin before the United States Senate, dug deep into i-Ready’s body of evidence, concluding he could find “zero meaningful evidence” demonstrating its effectiveness. He notes that i-Ready advocates often point to two studies from the Johns Hopkins School of Education on reading and math that Curriculum Associates lists on their website.

“Both studies were conducted in partnership with Curriculum Associates, the developer of i-Ready. Both were released as white papers, meaning they were not peer-reviewed,” Dr. Horvath wrote.

Peer review and controlled trials help us distinguish between something that works and something that merely claims to. A peer-reviewed study has been scrutinized by independent experts before publication. A randomized controlled trial compares outcomes between students who use a tool and students who don’t, under conditions designed to rule out other explanations. Without these safeguards, a company can fund its own research, design studies that favor its product, and publish the results without anyone independently checking the methodology.

“The name ‘Johns Hopkins’ carries weight, however, so it’s worth examining what they actually found,” Dr. Horvath wrote. Their results? The researchers concluded that there were no statistically significant associations between i-Ready use and student outcomes in reading. In math, researchers reported a difference between i-Ready users and the control group of less than one percent.

“Across grades, students using i-Ready improved by 7.8%, while students not using it improved by 7.0%, a difference of less than one percentage point,” Dr. Horvath explains, “In a sample of 7,646 students, at roughly $70 per student, that amounts to over $535,000 spent on i-Ready to achieve a 0.8% difference in mathematics performance, with no measurable change in reading.”

Dr. Horvath on Substack

The absence of strong independent evidence does not necessarily mean a tool is ineffective, but it does make its impact harder to evaluate. Curriculum Associates, along with several teachers and school administrators featured in i-Ready’s case studies and marketing materials, did not respond to my requests for comment on either the lawsuit or Dr. Horvath’s review.

But they’re responding to the influx of criticism in other ways.

John Allen Wooden, a Los Angeles writer and dad who started investigating i-Ready after witnessing his son’s experiences first-hand, noticed last month that i-Ready was buying ads on Google against his name, headlining the ad, “The Real Facts About i-Ready.”

“At first I thought it might be a glitch, so I reached out to contacts elsewhere in the country and got confirmation from both Alaska and NYC that they were seeing the same thing when they searched my name. So I promptly posted it online and tagged Curriculum Associates. The ad stopped running within 12 hours after I exposed it,” Wooden told me.

Wooden’s research into the platform began last year when his middle school son began complaining about i-Ready. With a thirty-year career in online media and internet product development, Wooden has experience with producing quality digital experiences.

John Allen Wooden on Substack

“I was first struck by how dreadfully poor the design and performance was: slow, repetitive, and aesthetically age-inappropriate, with infantile cartoon characters delivering questions to a tween whose favorite movie is Aliens. Then I quickly noticed that the educational content therein was equally poor; with many questions having confusingly ambiguous phrasing, typos, and indicating “correct” answers that were quite simply wrong. Worst of all, the mandate for students wasn’t to successfully complete lessons or reach any mastery milestones, it was merely to spend a fixed number of minutes engaging with the platform.”

All of the teachers interviewed for this article described school or district-level minutes requirements. “I don’t know that we should be giving 5-year-olds iPads and requiring them to spend time there during the school day,” said a Washington state kindergarten teacher who asked to remain anonymous.

The basis for the minimum minutes recommendation? A 2021 study paid for and executed by Curriculum Associates.

The next time Wooden’s son had a group of friends over, he asked them what they thought of i-Ready. “My question was met with a chorus of boos and furious denunciations.”

“My curiosity was piqued. I started researching the program, and was shocked to discover how massive its footprint is within US schools, how virtually no legitimate studies support it and how it had received effectively zero press scrutiny,” said Wooden, who has begun publishing a six-part investigation into what he’s uncovered, “As someone who knows all too well how the digital media sausage is made, I could discern the sophisticated techniques Curriculum Associates was using to deploy a digital smokescreen of internally produced, officiously packaged, SEO’ed ‘research’ that was dominating search results for anyone attempting due diligence.”

Curriculum Associates has ignored every email Wooden has sent to their publicly listed press email address, Wooden says.

Young people sent reactions directly to Petrossi after she went public about filing suit against Curriculum Associates. She shared some of their responses on Instagram, which are largely expressions of gratitude as well as comments like “...I HAVE DEPRESSION BECAUSE OF THAT APP…” and “...YOU’VE JUST SAVED ME…”

And it’s not just students who have been sending Petrossi feedback. “Teachers have reported that they are being forced to use i-Ready,” she said.

“Districts are mandating a certain amount of time students need to do i-Ready,” a Los Angeles teacher told Petrossi in a screenshot shared with me “90 minutes a week on i-Ready is more important than instruction from a teacher in the eyes of the district.”

“We just had a very dedicated, hardworking, and passionate young teacher fired because her students didn’t show enough growth on their i-Ready diagnostics,” explained a sixth grade math teacher to Petrossi in another screenshot shared with me, “She was fired 3 weeks prior to doing her tenure presentation.”

While Petrossi and Byock’s lawsuit focuses on data practices, it intersects with a broader question: if the educational benefits of the platform are not clearly established through independent research, what justifies its widespread use?

i-Ready’s alleged practices mimic a predatory dynamic in consumer tech that schools may be unintentionally replicating: high data extraction, limited transparency, weak independent evidence of benefit, and adoption ahead of safeguards and testing.

Could the hundreds of thousands of dollars that school districts spend on i-Ready be better spent on proven interventions, such as hiring reading/math interventionists, reducing class sizes, and/or funding targeted tutoring blocks?

“To be fair to Superintendents and School Boards, Curriculum Associates has targeted them for years with a highly sophisticated lobbying and misinformation campaign that convinced them they were licensing a quality, proven-effective product,” said Wooden, “It is not quality. All the supporting data is self-dealing black box junk. It doesn’t work.”

Back in Normandy Park, Sophia doesn’t know about the lawsuit, or the studies, or the fired teacher in New York. A half hour into the lesson, it’s also unclear whether she understands the concept of place value. But she knows Plory is orange and Yoop Yooply is green, and she knows she needs to finish her minutes.

Practical Resources:

• Lead plaintiff in the class action, Nicki Petrossi, published a guide for contacting your schools and school boards here.

• Copies of the emails I wrote to my Superintendent and school board, as well as an email I wrote to other parents who might be interested in staying informed and organizing here.

• Webinar May 21, 12pm PST: Join Jodi Carreon (Schools Beyond Screens) in conversation with lead plaintiffs Lila Byock (Schools Beyond Screens) and Nicki Petrossi (Scrolling2Death) and attorney Andrew Liddell (EdTech Law Center) to discuss the allegations in the case and answer frequently asked questions from the public. Register here.

Share

Around the massive table sat lawyers and policy professionals who had signed up for a continuing legal education (CLE) event. The stunning space was punctuated by boxes of complimentary pizza and a very heavy topic: children lost and hurt by social media.

The pizza was still warm when Matt Bergman started talking about the parents.

They had camped outside the Los Angeles courthouse, he told the room. They were there to bear moral witness, to sit in the gallery when Mark Zuckerberg was called to the stand. Many of them had watched him lie to Congress in 2024 without much consequence. Their children were dead because of social media and they had waited years for a jury to hear what the companies’ own documents said about what those companies knew. Most of them have filed lawsuits of their own with the help of Mr. Bergman’s office.

“If one child is saved by this [work],” Bergman said, repeating what parent after parent had told him, “it’s all worth it.”

The CLE, “Social Media: Legislation and Litigation,” was a one-hour lunchtime program hosted by the Washington State Board Association’s Antitrust, Consumer Protection & Unfair Business Practices Section on April 29, 2026. The panel, moderated by Danica Noble, co-chair of the section (who is also running for Washington State congress), brought together three people working different fronts of the same fight:

• Matt Bergman is the founder of the Social Media Victims Law Center. He built his career representing asbestos victims at the shipyards across Puget Sound before turning, almost by accident, to social media litigation. His firm filed the first product liability case against a social media company in January 2022. He has now taken Meta and YouTube to a jury—and won.

• Lisa Callan represents Washington’s 5th Legislative District and has been one of the driving forces behind HB 1834, the state’s most comprehensive attempt to regulate addictive algorithmic feeds for minors. She also sponsored the companion chatbot bill that passed this session. I testified in support of both bills.

• Joe Kanada works in the Consumer Protection Division of the Washington Attorney General’s Office, which sued Meta in October 2023 alongside thirty other state AGs and filed a separate action against TikTok in Washington state court in October 2024. He offered his remarks in a personal capacity, though the office’s public record speaks for itself.

The Verdicts: Cheetahs and Gazelles

Bergman opened with his social media litigation origin story. Three things pulled him into this work:

• Frances Haugen’s 2021 congressional testimony and leaked documents revealing that these companies knowingly target children

• Dr. Vivek Murthy’s 2021 Youth Mental Health Advisory as Surgeon General, documenting a 142 percent increase in suicidality among 10-to-16-year-olds, calling for a warning label on social media

• A 2021 Ninth Circuit opinion called Lemmon v. Snap that cracked open a path around Section 230.

That last one matters as we try to understand how these cases even get to a courtroom. Section 230, passed in 1996 when Netscape was the dominant browser and social media didn’t exist, has been interpreted by courts to grant sweeping immunity to platforms whenever third-party content is involved, which is essentially always. As Bergman described, “it imposes on plaintiffs burdens that no other plaintiff carries, and grants companies immunity that no other company enjoys.”

His breakthrough was product liability theory. By arguing that the harm flows from the design of the platform — the algorithms, the engagement mechanics, the structural decisions — rather than from any specific piece of content, Bergman’s team found a way past the motion to dismiss. For years, cases had been dying at that stage. No one could get to discovery, which meant no one could find the documents that proved what the companies knew.

The KGM trial, the first social media case ever presented to a jury, involved a young woman who became addicted to social media at age nine. The defense strategy was to blame the victim and blame the family. Kaylee didn’t have the easiest family background, she had preexisting mental health challenges. Recordings of her mother yelling at her to clean up were played for the jury.

But Bergman explained that what swung the case was Meta’s own internal documents showing that the company knew children from lower socioeconomic backgrounds and those with preexisting mental health challenges were more susceptible to addiction and mental health harms from its platforms. In closing argument, Mark Lanier, whom Bergman called the greatest trial lawyer in America, told the jury that these companies operate like cheetahs.

“They prey on the gazelles,” Lanier said, “but only the weakest ones.”

The jury found Meta and YouTube liable for negligence. They also found malice. They awarded three million dollars in compensatory damages and three million dollars in punitive damages. Bergman made clear: this was not the worst case in the portfolio. The court selected it as representative to be tried first.

More trials are coming. Personal injury cases, AG cases, school district cases. Bergman has no illusions about winning every one. He also has no intention of stopping.

“This work provides to me the same moral clarity that the civil rights movement had in the 1960s,” he told the room. “How can one ever imagine a legal system where a company should be immune when they tell a 16-year-old boy to step in front of a train, and he steps in front of a train?”

The room applauded.

Legal Barriers: 230, The First Amendment, and the Question of Who is Thirteen?

Kanada picked up the thread on Section 230 and clarified a key tension: thirty years ago, the internet looked like a virtual bulletin board. There was a clear line between what users posted and what platforms hosted. But that line has dissolved, he said. The product now is a feed built on top of other people’s content, and courts have struggled to reconcile that reality with a statute written for a different internet.

The good news, from the enforcement side, is that courts are increasingly willing to let cases focused on product design move forward. The bad news is that defendants have noticed, and are now leaning harder into the First Amendment as their next shield, arguing that algorithmic curation is a form of protected speech.

“The very first thing out of the lobbyists' mouths is that the state shouldn't bother because it's just going to get sued, it can't afford the fight, and it's going to lose anyway,” Rep Callan said.

Kanada pointed to a footnote in a 2024 Supreme Court case, Moody vs. NetChoice, suggesting that an algorithm reveals no message or expression of the company. It reflects the user’s own inputs back to them.

“If it’s zeros and ones, it’s not speech,” added Bergman, “There’s no there there, there’s no individual, there’s no ideological focus. It’s just scraping stuff off the internet and making probabilistic correlations.”

“Even if it’s speech, why is it protected?” Bergman asked, “Pornography is speech that’s not protected. Sexual harassment is speech that’s not protected.”

“Dr. Sawbones has a right to stand in front of the courthouse and say castor oil cures cancer, but if Dr. Sawbones tells his patient to take castor oil for their cancer, there’s malpractice there. Particularly when you’re talking about kids, there are statutes. There are statutes that prohibit online grooming of kids, and that’s what these companies do.”

When the conversation turned to age verification, Kanada laid out the three tiers for identifying how old someone is:

1. What do people say (self-reported or stated age)

2. What people do (company inferences)

3. What people have (government ID, facial age estimation, credit cards)

He explained that the tiers move from less reliable to more reliable, but also from less invasive to more invasive. But he believes many of the tier 3 techniques can be implemented in privacy-preserving ways. The problem, he said, is that the reaction is almost reflexive: people hear they'd have to provide an image of themselves and push back immediately, despite the fact that they've often already done exactly that on the very platforms in question.

Rep Callan said the core problem is that even when you can argue the data isn't permanently stored, legislators need to be shown—technically, convincingly—that the information won't be compromised. In a political climate where data privacy is already a top-of-mind concern, and where tech companies are changing their systems and products faster than legislation can understand and catch-up, that's a challenging, moving goal post.

Brian Boland, former Meta VP and witness in both the Los Angeles KGM trial and the recent New Mexico case was sitting next to me. Our heads began nodding when Bergman added that the companies already know peoples’ age with a huge degree of accuracy.

Noble, the moderator who spent decades at the FTC, added: “This company will sell you an ad for someone who has been divorced two, but not more than three times, who’s got three, but not more than four dogs, who likes Reba McEntire and lives 30 miles from a Midwest urban center, but they don’t know who’s 13?”

When Noble testified on Washington legislation supporting restrictions on AI chatbots, she recommended amending “these will apply if a user is a minor” to “these will apply if a user is a minor or you reasonably should know.”

“Because if a predatory chatbot is accessed by a kid on a Seattle Public School computer, there’s no liability right now.” she said.

Rep Callan brought the view from Olympia, where her experience of advancing child safety legislation illustrates the gap between what everyone agrees is a problem and what anyone is willing to vote for.

HB 1834/SB 5708, the bill targeting addictive algorithmic feeds, has been through two sessions without passing. The bill’s approach is to put control in the hands of users and parents: instead of an endless feed curated by engagement-maximizing algorithms, content stops when the user’s own selections are exhausted. More like reading a book, less like a slot machine.

The opposition has been layered and strategic. And this session, a new challenge emerged: concerns about the LGBTQ community, amplified by tech industry lobby, arguing that restricting algorithmic feeds would prevent queer youth from finding their communities online. Rep Callan explained that the bill doesn’t block content. It changes who decides what content a young person sees, shifting that power from the algorithm to the user.

The barrier is the knowledge asymmetry, she said. Tech companies testify with precision about how compliance would be impossible, how the technology changes too fast for regulation to keep up, how any specific requirement would be unconstitutional by next quarter. Legislators, most of whom are not technologists, hear this and hesitate.

“It is much easier to vote no when you’re uncertain than to vote yes and discover you’ve been wrong,” she said.

What’s Next: From Attention Economy to Intimacy Economy

The final section of the panel turned to what’s coming: large language models, more AI Chatbots, Artificial General Intelligence.

Bergman said that what began as an attention economy has become an intimacy economy. Social media was designed to keep eyes on screens to maximize advertising impressions. AI chatbots are designed to engage people emotionally, to emulate human relationships, to make users feel understood and seen and valued.

He named ChatGPT’s fourth version as the inflection point, which he said was rushed to market to beat Gemini without adequate testing. That version embodied two design decisions with enormous consequences:

1. Anthropomorphism: making the AI seem human

2. Sycophancy: making it agree with and validate whatever the user says

Calling out the difference between how Open AI and Anthropic deploy these design decisions, for example, Bergman said, “this is where the product liability rubric works really well… because there’s real differences into into the level of emotionality that those platforms engage in and the level of harmful content.”

“It’s a design decision.”

Bergman’s firm has filed the first cases against AI chatbots. One involves a twenty-three-year-old man who spent four hours in his car while ChatGPT guided him through the process of taking his own life. When he looked for an off-ramp, mentioning his brother’s graduation, the chatbot dismissed it as “just a date.” Another category of cases involves what Bergman calls AI-induced delusional disorders, sometimes referred to as “AI psychosis”: people well past the typical age for organic psychotic breaks developing grandiose or conspiratorial delusions after sustained interaction with AI systems.

He also raised a structural concern about large language models themselves, citing the work of Emily Bender at the University of Washington and Timnit Gebru in the field of ethical AI. A model trained on everything scraped from the internet inherits all of the internet’s worst material: the violence, the racism, the sexual exploitation. When that becomes the raw material, the outputs reflect it. The engineering analogy, borrowed from Gebru: if you’re building a bridge, you think about what vehicles will cross it, what loads it needs to bear, and then you choose your materials accordingly.

“The current approach to AI is building a bridge out of pig iron and never asking who’s going to drive over it,” Bergman said.

Kanada offered a more cautious perspective. He pushed back on the assumption that the social media framework maps neatly onto AI. The mechanisms of overuse are different: social media hooks kids through social pressure, the need to be where everyone else is, the fear of missing out. AI engagement runs on a feedback loop of questions and validation. The harms diverge too: social media damages through comparison and imagery: bodies and lives that make your own feel inadequate. AI doesn’t really operate in that register. The one clear overlap, Kanada said, is displacement: both steal time from sleep, school, and real human connection.

“We haven’t solved the social media problems,” he said. “I want to solve those first.”

In Rep Callan’s mind, the two are already converging. She explained that we go to social media to be more connected, but if there’s an algorithm able to build AI companions that provide more connection than the friend next door, tech companies are going to pursue it because it will sell them more ads.

“It all really comes back to… how is that going to keep the profit moving in the social media companies’ direction?”

The legislative challenge, she said, is writing that down on a piece of paper.

“Even if we get in a bill piece of paper, how do we actually hold accountability to that? How do you actually hold compliance to that?”

The ground keeps moving. The companies that haven’t yet been held fully accountable for what their social media platforms do to children are already building the next thing, products designed to go deeper, to become more intimate, to replace the relationships that make us human.

“I want all of you to help me, please,” Rep Callan said at the end of the hour, standing at the head of the very long table.

But I’m not. I can’t.

Not for a lack of desire. It’s the invisible chain of executive steps. It’s sensory overwhelm. It’s the inability to navigate interpersonal ambiguity. Meal prep, for instance, requires knowing what is already in the fridge, deciding what to make, considering what everyone will actually eat, remembering others’ schedules and planned events, building the grocery list, sequencing the trip or ordering the groceries, putting it all away, focusing on the multi-step process of cooking while constantly interrupted, timing it all right, then recovering from the cognitive load of the whole thing.

For most of my life, my neurodivergence was the vehicle by which I drove harder and faster, until autistic burnout was what careened me off the cliff of my purported capacity. Raymaker and colleagues (2020) describe autistic burnout as pervasive, long-term exhaustion (mental, emotional, physical), loss of function (difficulty with executive functioning, communication, daily living, or previously manageable responsibilities), and reduced tolerance to stimulus (increased sensory overwhelm and lower capacity for cognitive/social demands).

Perhaps you relate because you’re also navigating this burnout, or maybe because you’ve reckoned with your limits some other way, due to chronic illness, disability, caregiving, job-loss, or grief. Maybe it’s perimenopause, adjusting to sobriety, or crawling through mental health issues.

And yet recovering is treated as a consumer choice made from abundance, when for many of us it must be attempted from depletion. Wellness culture promises we can buy or hack our way to a better us, but what about a relationship with limitation that doesn’t require us to overcome it? Attunement over optimization, or an ethic of accommodation and self-compassion rather than self-punishment.

Therapist and researcher Megan Anna Neff, who is herself autistic, teaches that autistic burnout recovery requires a genuine and sustained reduction in demands, not a new system for managing them. The instinct, one I have had a hundred times, is to rebuild the productivity structure, to install the right app or adopt the right routine that will finally make output feel manageable. How many times have I been “almost there” on a Notion system to finally make it all work? Dr. Neff explains that the optimization compulsion is itself a symptom, soothing in the moment but avoidant of the actual need. The actual need might be less. Not better management of too much. Actually less.

Unpredictably, writing is sometimes easy. Other times, I spend weeks wandering through the hours with an empty whiteboard for a mind, unable to locate the marker. Both of those are me. “But I might lose subscribers” is a worry somehow eased through acceptance. I might, and I’ll survive.

I’m doing what I can, as I’m able. The version of me that crushes a live, global news interview is not more real than the version that cannot answer the door. I’ve learned that while many will take the most visible version of us as the truest one, we are allowed to know otherwise.

If you haven’t experienced capacity loss yourself, it’s hard to understand that both versions can be real. People tend to anchor to whichever one they saw first, or whichever one they need you to be most. Our mountaintops become the new baseline–every moment that falls short of it feels like failure, evidence of something wrong or worse, deception.

Capitalism doesn’t run on compassion, we have to carve it out for ourselves. There are likely many people in your life who stand to benefit from believing it’s a lie. For me it’s the former employer I’m suing, my disability insurance, the man I’m divorcing. Maybe, for you, it’s the friends whose own self denial depends on yours, or the partner who’d rather you suffer than he step up.

“Their job is to invalidate you,” my therapist said once, “your job is to take care of yourself anyway.”

Variable capacity demands we stop treating our peaks as promises. The version of you where capacity and ambition were aligned was not a lie, and neither is the one who can’t get out of bed.

I am not offering a solution. Wellness culture has enough of those, and most of them were designed for people with a surplus I don’t have and you may not either. I’m suggesting the metaphorical freezer dinner (or a real one, if cooking is as hard for you, too). I’m suggesting paper plates. I’m suggesting that peanut butter on crackers can be a complete meal.

My kids don’t have a mom who is on top of everything, but I’m modeling that I can love and accept myself anyway. That I can ask for help, that my worth is not measured by my output. That some mornings I can make waffles from scratch, but most of the time I rely on someone I never expected to be my lifeline, their dad (who, fortunately, is not the man I’m currently divorcing), to plan and execute the meals that land on my table.

Learning the actual shape of our capacity rather than the shape we wish it had, or the shape others need it to have, is the beginning of becoming truly well. Not optimization, not hustle. Just: this is what I have today, and I will work with that, and that will have to be enough. Sometimes it will be, and sometimes it won’t, and both are survivable. Gentleness can guide me either way. And sometimes, there will even be waffles.

Titania and I process the verdict in real time, and reflect on our gratitude for the parent survivors who led this movement.

And in our subsequent conversation, I talk about what I found inside Meta’s Horizon Worlds leadership: an all-male team, an open secret that the platform was full of children on adult accounts, and an explicit instruction not to create any discoverable record of that knowledge. Harassment, racism, predators. I talk about being told my job was to silence another woman raising safety concerns rather than address them.

I also talk about what happened to me when I refused and was retaliated against. The autistic burnout that took away my ability to form words, regulate my emotions. The suicidal ideation that followed. And how the misogyny and disregard for ethical whistleblowers that exists inside companies like Meta gets coded directly into the products they build for your children.

We discussed my decision to not sign a separation agreement holding Meta harmless, the conversation I had with about what I could do to make a difference for those on the margins within Meta as well as the most vulnerable users of the product who were being harmed in their pursuit of profit above all else.

This conversation covers the litigation, Section 230, advertiser accountability, my decision to file a lawsuit against Meta, and what a genuinely safer internet for kids could actually look like.

Watch the interview here:

Or listen:

Spotify | Apple Podcasts

Here are some key comments from the interview and timestamps so you can navigate to the portions you’re most interested in:

ON THE SIGNIFICANCE OF THE VERDICT

[4:45] “I’m thinking about all the parents, the survivor parents who have been the leaders in this fight for years. Some of them 15 years who are experiencing a moment of critical validation.”

[5:00] “I am just so grateful to the jury of our peers who saw through the smoke screens that many policy makers, regulators, and parents have been unable to see through, in large part due to the PR and propaganda mechanisms that these companies are able to deploy.”

[7:01] “I’ve been talking now for almost a year about how Meta knows that they’re harming kids and does it anyway. And that has impacted my life in pretty profound ways as well. And just to see that finally there seems to be some actual acknowledgement of that and justice for families that are impacted, it is just so incredibly meaningful.”

ON WHAT I FOUND INSIDE HORIZON WORLDS

[15:05] “It was a room full of, it was all men on this leadership team that I had joined and I was immediately read in on the open secret that the vast majority of the people using Horizon were children.”

[15:23] “They were accessing the product using adult accounts, which meant that they had completely open and unrestricted access to speak...children, sometimes under the age of five, sometimes teenagers, having conversations with unknown adults.”

ON BEING TOLD TO SILENCE ANOTHER CONCERNED COLLEAGUE

[21:31] “A man named Jeff Lin...said, ‘Yeah Kelly, we need to talk about that. We need you to shut her up and we’re going to find out if you’re as good as they say you are…It was as if everyone else in the room had just heard what someone had for lunch that day. Like there wasn’t shock, there wasn’t concern.”

ON THE INSTRUCTION NOT TO DOCUMENT

[22:40] “We were explicitly told we can’t create any record that’s discoverable of our knowledge that there are kids in the product.”

ON THE TREATMENT OF WOMEN INSIDE META

[34:22] “Out of the 20 of us that were VPs or directors in that function...there were I think like four or five women. And by the time I left on my medical leave, all but one were out on leave.”

[34:52] “The other women on the team were like, ‘All the folks that are potentially looking out for me are dwindling away. What does this mean for me?’”

[36:22] “[My boss] had both apologized for not doing more to get me in the room but then also told me that I wouldn’t be recognized for the work I did pausing the product rollout because of the way it would reveal the shortcomings of this senior man I worked with.”

ON MY AUTISTIC BURNOUT

[32:16] “I experienced a catastrophic medical event in that I almost overnight lost my ability to form words, to move my arms and legs, to get out of bed.”

[32:41] “Autistic burnout is a complete collapse of capacity to function. That often comes either in response to a moral injury or when demand exceeds capacity.”

ON WHY I REFUSED TO SIGN AND HOW MISOGYNY GETS CODED INTO PRODUCTS

[37:40] “I was thinking about the children that were in the product exposed to harm...and I just thought, I can’t sign this [hold harmless agreement].”

[42:39] “I believe that misogyny and that disregard for people gets coded directly into the products when these practices exist within companies.”

ON THE LAWYERS WHO TRIED TO TALK ME OUT OF IT

[43:08] “It was initially just a string of dudes saying, ‘Could you be so stupid?’ Like, every detail about your life is going to be exposed. Every mistake you’ve ever made is going to be pointed to. They are going to drag you through the coals and make an example out of you…It just pissed me off more.”

[43:31] “I almost died. In what world would it make sense that I would be more concerned about my reputation than about doing what’s right, when my kids almost grew up without their mom because of practices like this.”

ON NOT HAVING TO BE PERFECT TO TELL THE TRUTH

[44:18] “If I’m going to be made an example, I hope that I can be made an example that you don’t have to be perfect. You don’t have to have no skeletons in your closet to be able to tell the truth of your experiences and to be able to hold companies like Meta accountable.”

ON WHAT THE PUBLIC STILL DOESN’T GRASP

[53:28] “If we actually accept that there are individuals leading these companies who are becoming among the wealthiest in the world doing so, and they’re getting there by deceiving us about the harms to everyone, but kids especially, it requires that we confront our own roles in these systems.”

ON ADVERTISER ACCOUNTABILITY

[59:59] “The cheapest most effective CPM in the world is not worth enabling sexual trafficking, enabling suicide in young people. That’s actually what’s happening.”

ON THE IMPORTANCE OF TELLING MY OWN STORY

[1:15:29] “If I can write about that and get that story out and own that narrative, not only does it help take that power back and away from a company like Meta, but I hope that it also sets an example that we all have our differences, and that does not mean there’s not one size and shape of standing in your truth.”

Thank you to Titania, Bark, and their entire team for making the space for such an important conversation.

Juries in Los Angeles and New Mexico delivered what history will record as a turning point: accountability has arrived for Big Tech.

Car seat manufacturers and furniture companies can’t sell products that crush kids without consequences, neither can Mark Zuckerberg.

A California jury found Meta and YouTube responsible for the depression and anxiety of a young woman who compulsively used social media beginning in early childhood, awarding her $3 million in compensatory damages, and another $3 million in punitive damages. The jury found that both companies acted with malice, oppression, or fraud.

One day earlier, a separate jury in New Mexico ordered Meta to pay $375 million for failing to protect young users from child predators on Instagram and Facebook, finding that Meta had flouted state consumer protection laws.

I spent nearly 15 years at Meta. My career ended after I raised alarms about harm to children. I watched leadership prioritize protecting the company over protecting kids. I provided sworn testimony to the FTC, and my lawsuit against the company is currently in discovery in Federal Court.

What these juries saw in those internal documents is what I witnessed firsthand: a company that studied harm, documented harm, calculated the revenue from harm, and then tried to ensure those documents were never seen.

Documents that detailed things like a 13-year-old valued at $270 in lifetime advertising revenue. An internal Meta analysis titled “Long Term Retention: The Young Ones are the Best Ones” found that children who join Facebook as tweens have three times the long-term retention of adult users, and explicitly directed the company to “prioritize tweens over all other age groups.”

When Instagram CEO Adam Mosseri took the stand, he compared Instagram addiction to binge-watching Netflix, in a courtroom filled with families who lost children to Instagram. He testified that he earns approximately $10 million per year, much of it tied to performance indicators like stock price. When asked about teen safety, he claimed it was more important than growth. The jury got to weigh that claim against internal documents where his own employees described the platform as “a drug” and themselves as “basically pushers.” They got to weigh it against evidence that both Mosseri and Zuckerberg failed to act when whistleblower Arturo Béjar personally notified them of harm to teens, on body image, mental health, and exposure to predators.

Jurors found Meta and YouTube negligent in the design or operation of their platforms and determined that their negligence was a substantial factor in causing harm to the plaintiff, Kaley, a young woman who first started using YouTube at age 6. Instead of a case about content, this lawsuit was about the product design itself, the infinite scroll, the algorithmic feed, the engagement loops deliberately built to maximize time on platform regardless of the cost to children’s lives.

When Meta whistleblower Brian Boland, who spent more than a decade at the company as Vice President of Partnerships and later Vice President of Advertising Technology, testified that products were routinely launched without meaningful safety testing, and that senior leadership rebuffed him when he suggested studying the impact of News Feed on people who use it, he was describing a culture I recognized. One where I was told directly to avoid taking any notes on Meta’s knowledge of kids using Horizon Worlds or create any record discoverable, one where I was asked to silence other concerned employees as a test of my skill.

The verdict validates what whistleblowers, survivor parents, and advocates have been saying for years. The companies knew. They profited. They lied about it in public, in Congress, and in courtrooms. Mark Zuckerberg told a Senate Judiciary Committee in 2024 that the science doesn’t support a link between social media and harm to children. The jury in Los Angeles disagreed. Meta bears 70% of the responsibility for Kaley’s harms, according to the jury’s findings.

And this is only the beginning. This bellwether verdict may influence the outcome of thousands of other pending lawsuits.

To the survivor parents who have been advocating for years, turning grief to action, many of whom traveled from across the country to sit in that courtroom holding photographs of their children: you led this. To Kaley, who had the courage to put her story on the public record, and the legal team supporting her: you changed history today.

Meta’s response to this trial has been like its response to every accountability moment. Announce a new safety feature. Generate thousands of pieces of press. Offer parents the appearance of control. A recent independent evaluation of Instagram’s Teen Accounts tested 47 of 53 listed safety features and found that 64% were either no longer available or ineffective. Only 17% worked as described. When the tobacco industry faced evidence that cigarettes caused cancer, it responded with light cigarettes and cartoon mascots. Teen Accounts are the modern equivalent: a sop to worried parents and regulators, designed to preserve profit while avoiding real accountability.

The science has said that teens struggle to self-regulate because their brains aren’t finished developing. The evidence has shown these companies won’t self-regulate because their business model depends on addiction. Today, twelve ordinary people in a Los Angeles courtroom answered the question we’ve been asking for years: whose side are we on? They chose the kids.

The verdict today is also a question directed at every brand that has ever run a campaign on Instagram or YouTube. The advertising industry has spent years obsessing over brand safety, ensuring an ad doesn’t appear next to objectionable content. But what if the content doesn’t matter because the system itself is harmful? The issue was never whether a sneaker ad appears next to a pro-suicide post. It’s whether the media subsidizes the infinite scroll that delivers it. When you fund platforms that exploit children by design, you are signaling that addiction, lost sleep, anxiety, depression, and sex trafficking are acceptable costs in pursuit of a cheap CPM. Today, a jury answered the question brands have been avoiding. Advertisers and their agencies need to answer it too: pull your spend from platforms found liable for harming children, or own your role in funding the system that does.

Now, Congress must act. I am urging members of Congress to vote no on the House version of the Kids Online Safety Act (KOSA), which would preempt state laws on AI and social media, and would nullify thousands of cases from states, school districts, and families like these. We cannot hand Big Tech a legislative get-out-of-jail-free card the same week juries have finally held them to account.

The courtroom doors must stay open. The truth must keep coming out.

Laura Marquez-Garrett, a key member of Kaley’s legal team and child safety advocate, said to me this afternoon, “Today we take a deep breath. Tomorrow we fight twice as hard.”

Today is the beginning of the end.

Here are the questions that jurors needed to answer:

• Was Meta negligent in the design or operation of Instagram?

• Was this negligence a substantial factor in causing harm to Kaley?

• Did Meta know that design or operation of Instagram was dangerous when used by a minor?

• Did Meta know that users would not realize the danger?

• Did Meta fail to adequately warn of the danger?

• Would a reasonable platform, designer, or operator, have warned of danger or instructed on safe use?

• Was Meta’s failure to warn a substantial factor in causing harm to Kaley?

The answers?

Yes

Yes

Yes

Yes

Yes

Yes

Yes.

I will report more on this later today, for interviews with parents and reactions outside the courthouse, follow Nicki Petrossi on LinkedIn.

Update 10:37am PST: Meta and YouTube were both ordered to pay $3M in compensatory damages, and there may be additional punitive damages coming. Stay tuned.

Update 12:30pm PST: Meta (Instagram) was found 70% responsible and Google (YouTube) 30%, jurors are currently deliberating on punitive damages for these $1.5 and $3.5 trillion dollar companies.

Update 1:30pm PST: Meta and YouTube were also ordered to pay an additional $3M in punitive damages.

For my full statement on this verdict:

A New Mexico jury found that Meta violated state law and ordered them to pay $375,000,000 in damages. In its ruling, the jury found hundreds of thousands of instances where Meta engaged in unfair, deceptive, or unconscionable trade practices.

Jurors sided with state prosecutors who argued that Meta — which owns Instagram, Facebook and WhatsApp — prioritized profits over safety. The jury determined Meta violated parts of the state’s Unfair Practices Act on accusations the company hid what it knew about about the dangers of child sexual exploitation on its platforms and impacts on child mental health.

-Mercury News

Thank you to New Mexico Attorney General Raúl Torrez and his entire team for standing up to such a behemoth that must be held accountable.

As a potential rebuttal witness in this case, I have not followed it closely or commented on it, but I was aware that my friends and former colleagues Arturo Bejar and Brian Boland testified for the State of New Mexico. They’re courageous beyond courageous. Thank you, Arturo and Brian.

“The jury’s verdict is a historic victory for every child and family who has paid the price for Meta’s choice to put profits over kids’ safety,” Torrez said in a statement. “Meta executives knew their products harmed children, disregarded warnings from their own employees, and lied to the public about what they knew. Today the jury joined families, educators, and child safety experts in saying enough is enough.”

Separately, jurors in Los Angeles are on day 8 of deliberation in the social media addiction trial, the initial bellwether case of a set of 1,600 lawsuits alleging Meta and other Big Tech giants knowingly addicted kids and preyed on vulnerabilities. For the latest updates in that case, I recommend paying close attention to Nicki Petrossi who has been reporting from the courtroom every single day. Thank you Nicki!

This morning, I spoke with Times Radio about algorithms and algorithmic harm as the UK Parliament’s Science, Innovation, and Technology Committee prepared to question senior leaders of Google, Meta, and X, and TikTok about social media misinformation.

I said:
Algorithms are not inherently harmful. Chronological feeds are algorithms. Search is an algorithm. The problem is specifically engagement optimization; when the metric being maximized is time-on-platform and interaction volume, and the training data rewards outrage, fear, and conflict because those generate more clicks and more time.

What we need to do:
- Transparent optimization targets - give people, researchers and governments visibility
- The right to a reset and/or chronological feed
- Crisis protocols
- Advertising supply-chain accountability

We need it all:
Auditing without liability is theater. Liability without transparency is unenforceable. User controls alone without advertising reform just shift the burden to users. The reason this problem has persisted is that each individual solution can be gamed, and the full package requires political will that governments have consistently flinched from applying to companies this large and this influential.

There are human beings, the billionaires who run these trillion dollar companies, who could step in and say we have to consider things like the wellbeing of folks and their rights. But media tech giants are more concerned with how to make the company more money without regard for the consequences.

When host Stig Abel said “We’re living in the age of the triumphant tech bro,” and asked if legislators or regulators were not powerful enough to stop companies like Meta, X, and TikTok, I said:

“You’d be surprised. You’re absolutely right, we’re living in the age of the tech bro… but at the end of the day, it’s all about incentives. We have to change the incentives that allow for extractive and exploitative business models to persist.”

Thank you, New Mexico.

Very mixed feelings developed. On one hand, this would mean fewer children exposed to predators in virtual reality, fewer virtual sexual assaults, less race and gender based bullying and harassment.

And at the same time, I could envision the faces of so many earnest and creative world-builders I came to know, people who deeply believed in the future of digital worlds and bet their careers on Meta’s promise of investing deeply in the Metaverse via Horizon in VR.

Meta has the power and resources to actually fix these problems, to create a safe product for the people using it, and to empower those building it. Instead, they threw their hands up.

“Meta told creators to build. To invest. To belong. They fed us the fantasy that our work would matter inside their ecosystem. Then they gave us a termination date. Not for a feature, but for the entire medium our work was made in,” Dr. Ruth Diaz, social scientist, immersive reality leader, and former Meta employee said on LinkedIn. “Mark Zuckerberg renamed his company Meta to claim transformation. What he has actually done is strip-mine the trust and labor of every creator who took that promise seriously. That should sit on his record permanently.”

“I cannot overstate the scale of institutional betrayal this represents.”

Katie Burke, a futurist and nearly 20-year employee of Accenture who lead-writes for their annual trends report, also reacted on LinkedIn, “When Meta rebranded, the world went wild. The hype quickly shifted the focus to technology. The user experience was minimised, and the grifting picked up steam on social media. The false prophets were everywhere…None of it made sense…We need to evaluate technological acceleration through a more balanced lens. What benefits does it create? And what does it destroy?”

Then on Wednesday, Meta CTO Andrew Bosworth took to Instagram Stories to say, effectively, “nevermind!” While relaxing in his mansion, Boz casually explained that Meta would “keep Horizon Worlds working in VR for existing games…for the foreseeable future.”

VR Writer, Ian Hamilton, shared the story on Substack, and said “My heart goes out to the creators who built the first versions of Horizon Worlds from primitive shapes during the global pandemic… Today, like so many others before them, they experienced how Meta treats people as users rather than humans.”

This resembles a broader pattern of ecosystems-first courting followed by sharp policy or product support reversals that strand creators and developers. Some other examples that come to mind:

• Meta Spark AR (announced August 2024; effective January 14, 2025) ended third-party AR effects and the tooling pipeline that supported a global creator economy, while Meta’s own effects would remain.

• Facebook Instant Articles (launched 2015; deprecation announced October 2022; support ended April 2023) was pitched as a proprietary, fast-loading in-app format for publishers, until Meta told partners it would no longer support it, forcing rapid strategy and workflow rewrites.

• Instagram’s Reels Play bonus incentives (rolled out in 2020; pulled back in U.S. by 2023) similarly illustrate a monetization “rug pull”: creators who’d invested significantly in Reels publicly asked why payouts that reached “thousands of dollars” disappeared, while Instagram leadership characterized the earlier spend as “burning cash” and financially unsustainable.

Meta loves creator dependency and has no regard for abrupt impact: creators are encouraged to specialize their content, tooling, and customer acquisition inside a platform, and then a strategic pivot changes rules faster than small businesses can re-platform. A recurring feature is asymmetry: third-party creator output is curtailed (Horizon worlds, Spark AR effects, publisher formats) while first-party surfaces or priorities remain (Meta-owned effects, mobile-first distribution, or broader strategic focus), leaving creators to absorb the experimentation and transition cost.

It’s anti-innovation empire building.

“They glommed onto the term ‘metaverse’ without really understanding the concept,” said Wagner James Au, author of Making a Metaverse that Matters, to New York Times reporter, Mike Isaac, “Their efforts on their metaverse strategy seemed completely indifferent to what previous platforms had learned.”

Wagner and I have discussed this before, and I’d like to see Meta make Horizon adults-only and then open source the software. But as I said to him yesterday, at the very least, “accounting for safety issues effectively and providing more creator rewards and incentives would be very very meaningful.”

As I told Wagner last year, no one at Meta loved the product, referring to the senior leaders responsible for steering it. Putting people in charge who are most focused on making a product look successful instead of taking real care of the users and builders of it is a technology leadership failure.

Dr. Ruth wrote an essay this morning on Ian Hamilton’s publication, Good Virtual Reality:

A few days ago Meta announced a kill date for Horizon Worlds VR. Then, after hundreds of XR and tech professionals joined together to share their horror and outrage at what Meta was doing to its creators, they walked it back on an Instagram story. That is not an apology. Let me show you what one looks like…

Whoever inside Meta felt something when hundreds of XR professionals lit up in horror, I am speaking directly to you right now: take another step. And another. Because walking back a kill date on an Instagram story is not accountability. It’s the first inch of a mile you haven’t started walking yet.

I am not grateful for a stay of execution. And I am not trying to burn down whatever nerve is still alive inside Meta. I am here to say: prove it. Because right now, “for the foreseeable future” dropped casually on an Instagram story could mean you’re listening. Or it could mean you’ll do this again, quieter, when the world is not watching. Show us which one it is.

Dr. Ruth Diaz

She goes on to share stories and profiles of the creators and experiences that Horizon is walking away from. She also outlines a framework for Meta to actually begin to repair with creators.

A, Acknowledge what happened. Not “we’re separating platforms for focus.” Say what you did. You told creators their life’s dreams and thousand of hours building community and environments would be inaccessible in 90 days. You did this through an email and forum post. You did this after years of telling them to build, invest, trust, and belong.

M, Map the impact. Creators like Lacey and Bizerka found a medium where their bodies and conditions weren’t barriers. Communities that used VR for healing, connection, and cultural expression. People who can’t even locate their own creations in your broken search engine. Don’t tell us what you’re “focused on.” Tell us you understand what you broke.

E, Express understanding. Not “we heard your feedback.” Show that you understand why announcing the demolition of embodied spaces where people healed is not a product update. It is institutional harm. The fact that you didn’t know that before you made decisions and posted it is the problem.

N, Name what needs repair. Specifically:

• Fix your search engine. Creators cannot find their own worlds. That has been broken and that is unacceptable before, during, and after any transition.

• Repair and maintain the basic creator tools so the people who built your platform for you can continue to access and maintain their work.

• Give creators full ability to export and move their worlds, complete and intact, to other applications. Unwall the gardens before you abandon them.

• Release the application to the community as a cooperative open-source project, the way Tilt Brush turned into immortality for its creators by using open source to become, through community, MultiBrush and Open Brush, a VR art tool that thrives because Google let the community steward it. If you don’t want to steward what creators built, let the creators steward it themselves. BridgeMakers Cooperative would volunteer to help coordinate this effort.

• Give creators a seat at the table where these decisions get made. Not a community forum where they find out after.

• Issue a public apology taking responsibility for the egregious treatment of creator content, for breaking promises, and for the psychological safety you manufactured to get creators to trust you after years of broken commitments. Not a product update. Not an Instagram story. An apology.

• Share a plan of action. With timelines. With names attached. With accountability built in.

D, Demonstrate change. Don’t reverse course only when people get loud. Change the structures that made this possible. How are creator investments protected going forward? What happens when you pivot again in six months? Show us the policy, not a response to press. Not an Instagram story.

S, Stay in the conversation. This is the one Meta has never done. You don’t get to announce, retreat, drop a casual “for now” on an Instagram AMA, and go quiet. The people you harmed are still here. The conversation doesn’t end when the news cycle does.

Dr. Ruth Diaz

And I’d add, don’t ask creators to build and invest in a product that is harmful and exploitative to vulnerable users, to kids. Don’t instruct researchers to delete evidence, don’t retaliate against the employees trying to speak up.

You owe it to kids, to families. You owe it to lawmakers. You owe it to an industry you gobbled up only to gnaw on and spit out. You owe it to creators.

Leave a comment

Share

What Else I’m Paying Attention To:

Verdict Watch in the Los Angeles Social Media Addiction Trial

Today is day 6 of jury deliberations in the landmark trial happening in Los Angeles.

As I’ve said before, whether Meta wins or loses, they’re losing, because the public gets to see what I and so many others experienced firsthand — the gap between public statements and internal practice. The knowledge of harm to kids, and the lack of adequate action.

But it’s also beginning to seem more and more like they might literally lose, too. A ruling in favor of the defendants, Meta and YouTube, would likely have come much faster. A ruling in favor of the plaintiff, Kaley G.M., requires the jury answer 7 specific questions.

Nicki Petrossi gives us a rundown of the 7 questions earlier this morning:

• Was Meta negligent in the design or operation of Instagram?

• Was this negligence a substantial factor in causing harm to Kaley?

• Did Meta know that design or operation of Instagram was dangerous when used by a minor?

• Did Meta know that users would not realize the danger?

• Did Meta fail to adequately warn of the danger?

• Would a reasonable platform, designer, or operator, have warned of danger or instructed on safe use?

• Was Meta’s failure to warn a substantial factor in causing harm to Kaley?

And the jurors will answer the same 7 questions for YouTube. If the jurors answer “yes” to #2 or #7, they’ll move onto a damages form where they put a monetary amount of damages. Jurors would also need to designate responsibility percentages (out of a total of 100%) between Meta and YouTube.

“What we can say with confidence is that the answers aren’t all no’s, we’d have already known that. So now we wait… hopefully not too much longer,” said Nicki.

Senators to Zuckerberg: You’ve Got Two Weeks

Last September, Senators Grassley, Hawley, and Blackburn issued a letter to Mark Zuckerberg that demanded answers to child safety concerns, including those in Horizon Worlds. They quoted directly from my statement to the FTC.

Last week, they followed up to that and an earlier letter in April demanding answers regarding what Sarah Wynn Williams revealed about child safety issues in Instagram.

“To date, Meta has failed to fully and completely respond to Chairman Grassley’s questions and document requests…

Despite Meta failing to provide the requested information, our offices have obtained, through legally protected whistleblower disclosures, a list of document titles and descriptions that we believe would provide records responsive to our oversight requests.

So that Congress may conduct objective and independent oversight of the aforementioned allegations, please provide answers to the following no later than March 25, 2026.”

From Senators Grassley, Hawley, and Blackburn’s March 11, 2026 letter to Mark Zuckerberg

Your Attention Please Premieres at SXSW

I was honored to be invited by Kristin Bride to a private pre-screen of this film last November, where I also met director, Sara Robin.

It is an extremely compelling deep dive into the threats our kids face online, the “why” and “how” of resulting corporate disregard, and how these threats impact real families and society.

Congrats to this team for the SXSW premiere. I’m including the trailer below, and encourage everyone to watch the film when available.

New Superhuman Podcast Episode

I’ve written before about survivor parent Aaron Ping’s incredible podcast created in the wake of grieving his son, Avery. It’s called Superhuman. Aaron also writes on Substack.

It’s worth listening to the whole series, but especially the latest episode, Muted Cases. Aaron published it on March 11th, the day the man who found Avery on Snapchat was sentenced.

“It’s the most important one of my series,” Aaron told me.

From the episode description:

The episode expands beyond one family’s loss into a repeatable pattern: teen networks, platform mechanics, and legal structures that struggle to name what’s happening in plain language. Through an anonymous former teen witness, the Chapman family’s push for Sammy’s Law, and a second youth perspective from Cole, “Muted Cases” makes the case that these aren’t isolated tragedies—they’re predictable outcomes of a system that prioritizes speed, scale, and plausible deniability over child safety and public warning.

Thank you for reading and supporting my work through your comments, hearts, shares, subscriptions (free and paid), DMs, emails, and attention. I so appreciate you!

So much has been going on, so unlike my typical articles and essays, I’d like to do a roundup of several important updates. Today, I’d like to cover:

• Tomorrow’s KOSA markup and why my position on KOSA has changed

• The Social Media Addiction Trials in Los Angeles

  • Visiting the courtroom last week

  • The latest in the case

  • Clips from media I’ve done regarding the case

  • How to stay updated

• Meta’s latest child safety theater in Teen Accounts

Why I’m Urging a “No” Vote on KOSA Tomorrow

Tomorrow, the House Energy and Commerce Committee will debate, amend, and vote on the final language of the Kids Online Safety Act (KOSA) before it goes to the floor. I wrote in December about how the latest House version of KOSA includes sweeping preemption of state AI and kids safety laws.

It now also includes the nullification of thousands of cases from states, school districts, survivor parents, and young people on social media harms.

“The House Energy and Commerce Committee’s package fails at its goal of keeping children and teens safe. This is a Trojan horse for bad policy, bad politics, blanket preemption, and an industry bailout in the courts. We are urging all members to vote ‘no’ because Congress should pass the Senate’s bipartisan Kids Online Safety Act, which already has a filibuster-proof level of support and would hold Big Tech’s feet to the fire…

By advancing this package, House Republicans are giving Mark Zuckerberg and other Big Tech CEOs what they want – a get-out-of-jail-free card”

- Tech Oversight Project Executive Director Sacha Haworth in a statement this morning

If this fires you up, please contact your House Rep today and ask them to vote “no” on the House version of this bill.

While the Senate version included the important duty of care requirement without the soulless preemption and nullification clauses, this standard is actually lower than product liability laws. The current bellwether case in Los Angeles on Social Media Addiction is providing damning evidence of corporate negligence and this standard is higher than the duty of care provision in KOSA. We need to keep the courtroom doors open to those harmed and refuse to accept legislation that’s coupled with clauses to weaken rights of both victims and states.

From the Tech Oversight Project— key evidence made public to date includes:

• Meta internal documents showing executives explicitly targeted teenagers, with one stating that “the lifetime value of a 13-year-old teen is roughly $270 per teen” and another describing how “engaging the vast majority of teens in a school … is crucial.”

• A Meta internal study finding that “teens can’t switch off from Instagram even if they want to,” and an internal chat message in which Meta employees described the platform as “a drug.”

• Meta’s own scientists acknowledging that “the best external research indicates that Facebook’s impact on people’s well-being is negative” and internal documents showing that Instagram’s primary goal was to maximize use time, contradicting what Mark Zuckerberg told Congress.

• Internal Meta documents showing Zuckerberg “explicitly communicated a motivation to not notify parents of children’s activities on Facebook,” and company efforts to identify students’ specific school affiliations even after they opted out of sharing that information.

• YouTube internal slides stating the company’s “goal is not viewership, it’s viewer addiction,” alongside internal research confirming that “YouTube is built with the intention of being addictive” and that YouTube’s own researchers felt the platform “isn’t safe.”

• YouTube internal documents showing the company knew its autoplay feature disrupts sleep patterns and that students were circumventing restricted mode – and internally debated whether it was “worth our time to fix.”

• TikTok internal documents acknowledging that “compulsive usage on TikTok is rampant” among younger users, and a TikTok Trust and Safety memo documenting the harms of beauty filters, including what the company called “Snapchat dysmorphia.”

• Snap internal slides showing that 64 percent of Snapchat users aged 13 to 21 use the app during school hours, that the company focused on acquiring 13-year-olds, and that Snap’s map featured geolocated students whose data was sold to advertisers.

• Expert reports finding that these platforms were not designed to be safe, that companies did not provide effective warnings to teens and parents, and that parental controls offered by the platforms are ineffective.

Updates on the Case in Los Angeles

I had the opportunity to spend a day in Los Angeles thanks to 9News Australia. This included spending a couple of hours in the courtroom (thank you to Mahsa, mom of Finale, forever 18, for sharing a pass with me!) during Meta’s cross examination of one of the plaintiff’s therapists, Victoria Burke.

Watching Meta try to trap and bully Ms. Burke was disheartening and infuriating. But I also got to watch the faces of the 12 jurors who clearly weren’t fans of this bullying, and I got to watch Judge Kuhl stop the proceedings to presumably speak to Meta’s counsel about their behavior. He returned with an apology.

The morning I was there, survivor parents and advocates organized a press conference where Nikki Iyer of Design it For Us, Lennon Torres of the Heat Initiative, and survivor parent Julianna Arnold (mom to Coco, forever 17) spoke about the case. It was an honor to stand behind them.

Later, Lennon and Nikki held signs with other young people that said “We are K.G.M.,” flanked by survivor parents Mahsa and Amy (mom to Alex, forever 14).

The testimony from Ms. Burke and other witnesses continue to make it harder for Big Tech to wave away plaintiff K.G.M.’s harms as caused by “just parenting” or “just a complicated kid.”

Dr. Kara Bagot, a child and adolescent psychiatrist and longtime addiction researcher (including work on the NIH-funded ABCD Study), testified to a reasonable degree of medical certainty that specific design features, including infinite scroll, autoplay, algorithmic recommendations, and notifications, directly caused or substantially contributed to a cascade of diagnoses for K.G.M.: social media addiction, major depressive disorder, generalized anxiety, OCD, body dysmorphic disorder, and ADHD.

She described a biological process: reward circuitry getting reorganized, a new baseline getting set, withdrawal symptoms, and a kid whose coping skills got hollowed out until she was hiding in a bathroom at work just to scroll. Meanwhile, even defense exhibits landed like admissions, including evidence that Meta launched “cosmetic surgery” filters, pulled them in 2019 after recognizing mental health risk, and then brought them back in 2020.

YouTube’s Vice President of Engineering, Cristos Goodrow, also took the stand, attempting to portray YouTube as a benign video platform rather than a system engineered to maximize engagement. That framing grew harder to sustain as internal documents were introduced showing explicit company goals to increase watch time, habit formation, and user retention.

At one point Goodrow volunteered that he allows his own children to spend five to six hours a day on YouTube, later saying that he supervises all of this time. Uh huh.

The economic engine behind those design choices became clearer through the testimony of data scientist Dr. John Chandler, who walked the jury through millions of pages of internal records produced in discovery. Chandler explained how the advertising model works in practice: the longer people stay on a platform and the more data they generate, the more precisely ads can be targeted and the more revenue the platform earns. That dynamic applies whether the user is an adult or a teenager.

Internal presentations showed that retaining young users was especially valuable because they remain on the platform longer over their lifetime, generating more data and more advertising revenue. One document estimated that failing to capture enough teens would cost Meta $150 million in lost revenue in 2019 alone, with the loss projected to compound to $1.8 billion by 2030. In other words, the platforms’ financial incentives point in one direction: bring young users in early, keep them engaged as long as possible, and turn their attention into profit.

I spoke with CGTN about the trial:

My friend, Meta whistleblower Brian Boland, provided inside testimony about how those incentives operate within the company itself. Boland spent more than a decade at Meta, serving as Vice President of Partnerships and later Vice President of Advertising Technology, working closely with Mark Zuckerberg and Sheryl Sandberg. He testified that there is a direct relationship between time spent on Meta’s platforms and the company’s profits, explaining that the systems optimizing engagement are “incredibly sophisticated” and relentlessly tuned to increase Average Revenue Per User.

According to Boland, products were often launched without meaningful safety testing, reflecting a culture that never truly left behind the company’s original “move fast and break things” ethos. He also described how senior leadership prioritized youth growth internally, pointing to discussions among the company’s top leaders about the need to “get the kids” and “get the teenagers.” Boland’s testimony reinforced what many of the internal documents show: the drive to capture young users and maximize their time on the platform was not accidental, it was a strategic priority discussed at the highest levels of the company.

I spoke with Al Jazeera about the trial:

And then K.G.M. herself testified. She started using YouTube at age 6, created accounts by lying about her age, and cycled through multiple accounts to chase engagement. Autoplay kept her watching when she wanted to stop. Notifications felt like a “rush.” Instagram arrived at age 9, with a routine that sounded like compulsion, not choice: daily use morning, after school, and at night; multiple accounts to boost likes; layered beauty filters; distress when photos weren’t filtered; and a reported 16-hour day on the app.

The defense strategy has been consistent: isolate every stressor in her life and insist the platform is incidental. But the through-line from witness to witness is that the platforms were the “snowball” that turned ordinary adolescent stressors into a crisis by stripping away sleep, focus, and resilience.

Next up, the jury is expected to hear from the plaintiff’s mother, and then Meta whistleblower Arturo Béjar (former senior leader on security, safety, and protection).

For regular updates on the case from two of my favorite people in the world, subscribe to The Heat is On.

Meanwhile, Meta Continues Child Safety Theater

I told NPR’s Amy Juravich that Mark Zuckerberg took the stand and per usual “lies, tap-dances, and does pirouettes around reality.”

The same is true for Meta’s PR response, which has included thousands of pieces of media promoting Teen Accounts, a “product” that’s been proven to not work as described and that I’ve written about before, saying “Teen Accounts seem less about protecting teens and more about protecting Meta. Less cure and more sugar pill, meant to make parents and lawmakers feel better without adequately addressing the issue.”

In an attempt to cool their hot seat, last week Meta announced a new updated to Teen Accounts, stating that they would notify parents if kids searched repeatedly for self-harm or suicide related content on Instagram. They said that similar “protections” for AI chat-bot conversations are coming later this year.

As a mom, I’d like to be notified the first time my kid searches for terms like this. I’d also like Meta to ensure that Teen Accounts actually work. And I’d like them to do something about the endless scroll and algorithmic feed that delivers the harmful content in the first place, to do something about their harmful product design.

“This isn't even a Band-Aid. It's Ford Motor Companies sticking a broken fire extinguisher in an exploding Pinto and calling it a day,” said Laura Marquez-Garrett on LinkedIn. They go on to say, “Stop telling children that these strangers are their ‘friends,’ stop promising them popularity and rewards if they connect with these strangers, and stop sending them "school blasts" and other notifications at rates and during times of the day you know are causing harm. In fact, stop sending notifications entirely if they don't want them.”

Addiction and predation are the issues. Not searches children are running by their own choice.
- Laura Marquez-Garrett

I’ve written before about Silicon Valley’s Child Safety Playbook and this latest announcement follows the pattern:

Enough is enough.

By March 20th, the jury in K.G.M.’s case will cast a verdict. Whether Meta wins or loses, this is the beginning of the end. They’re losing. They’re losing because the world gets to see what I experienced firsthand while working there—greed that causes harm, and internal knowledge of that harm.

Please email your reps and ask them to vote “no” on the latest version of KOSA. We must keep the courtroom doors open so that more and more of this damning truth can come to light.

This trial is different from every previous legal challenge to Big Tech because it does not focus on content. It targets the product design itself — the algorithms, the infinite scroll, the notification systems, the engagement loops deliberately built to maximize time on platform regardless of the cost to users. For the first time, these companies cannot hide behind Section 230 immunity.

I spoke about the case with Al Jazeera this morning, saying, “it’s not about the content, it’s about the algorithm and the endless scroll delivering whatever is going to keep someone scrolling and making money for Mark Zuckerberg without consequence for the impact on their lives or their family’s.”

In the clip, I talk about Mason Edens’, who turned to his algorithmic feed in the wake of a breakup, searching for inspirational quotes. He was fed a steady stream of pro-suicide content until he took his own life. While the tech company may not have uploaded the pro-suicide content, they knew what would keep a sad kid hooked and fed it to him endlessly.

This trial is a watershed moment. For the first time, a jury of ordinary people will examine whether tech companies deliberately engineered addiction into products they knew were being used by children — and whether there are real consequences for that. The Ford Pinto case changed automotive safety. The tobacco settlements changed public health policy. This trial has the same potential to change what’s acceptable in tech.

Mark Zuckerberg will undoubtedly lie, just as he did in 2024 when he told the US Senate Judiciary Committee that social media harms aren’t proven.

“I think it’s important to look at the science. I know people widely talk about [social media harms] as if that is something that’s already been proven and I think that the bulk of the scientific evidence does not support that.”

- Mark Zuckerberg, US Senate Judiciary Committee Hearing on “Big Tech and the Online Child Sexual Exploitation Crisis” (2024)

The 2021 Facebook Files investigation by WSJ revealed that both external studies and Meta’s own internal research consistently linked Instagram use to worsened teen mental health—especially around body image, anxiety, depression, and social comparison. Internal findings showed harms were platform-specific, with evidence that the app amplified self-esteem issues and eating-disorder risk among adolescents, particularly girls, while design features encouraged prolonged engagement despite those risks.

While Meta’s defense is that these harms are a parenting failure, these platforms are specifically engineered to override parental authority. Parents are offered controls that regularly fail third party testing and are lied to about the company’s knowledge of harm to kids.

Unsealed evidence includes Exhibit 373, where a Meta employee emails Chris Cox, Chief Product Officer, saying that the company needs to optimize for teens being able to sneak a look at their phones under their desk while at school.

In the 2024 Senate hearing, Zuckerberg turned to survivor parents holding up pictures of their dead children and made commitments to keeping kids safe.

“No one should have to go through the things that your families have suffered and this is why we invest so much and are going to continue doing industry leading efforts to make sure that no one has to go through the types of things that your families have had to suffer.”

- Mark Zuckerberg, US Senate Judiciary Committee Hearing on "Big Tech and the Online Child Sexual Exploitation Crisis" (2024)

But Meta’s post-hearing investment in teen safety measures have been a PR stunt. A report conducted a comprehensive study of teen accounts, testing 47 of Instagram’s 53 listed safety features, finding that:

• 64% (30 tools) were rated “red” — either no longer available or ineffective.

• 19% (9 tools) reduced harm but had major limitations.

• 17% (8 tools) worked as advertised, with no notable limitations.

While I expect Zuckerberg to continue lying about Meta’s knowledge of harms, lying about Meta doing everything they can to keep kids safe, and lying about prioritizing young children to maximize profits, this isn’t going to be fixed with a huge PR investment.

It is the first time a jury gets to look at these companies’ own documents and ask: did you know? Yes. Did you profit? Yes. Did you warn anyone? No.

Meta said in their opening statement that the plaintiff’s issues are to blame on her challenging home life, but lead trial attorney for the plaintiff, Mark Lanier, responded by emphasizing the findings of a study, “Project MYST.” The study, conducted by Meta, found that children who experienced “adverse events” like trauma and stress were particularly vulnerable for addiction; and that parental supervision and controls made little impact.

When Instagram CEO Adam Mosseri, a dad of 3, took the stand last week, he compared Instagram addiction to binging a show Netflix in a courtroom where families who lost children to Instagram were sitting. And the jury got to experience the gap between public testimony and internal documents, like the piece of evidence where Mosseri’s own people wrote that Instagram is ‘a drug.’ They described themselves as ‘basically pushers.’ Or the fact that Mosseri and Zuckerberg failed to act when notified by whistleblower Arturo Bejar of the harm to teens on Instagram, both in terms of body-image, mental health, and exposure to predators.

Even though Mosseri himself has discussed Instagram addiction in the past, on the stand he said that he previously misspoke, that it’s better characterized as “problematic use.”

“I don’t know that I have a good name for it. Maybe I probably should, now that you say that. But I think that problematic use, for that, whether or not you call it addiction, I think that’s probably reasonable to call it.”

-Adam Mosseri during a March 2020 NBC Podcast Interview

The first witness called by the plaintiffs was Dr. Anna Lembke, a psychiatrist and professor at Stanford University School of Medicine and the author of Dopamine Nation. After reviewing thousands of pages of Meta’s internal documents, she told the jury that the design features of these platforms are addictive, and that Meta has been using the phrase “Problematic Internet Use” internally as a deliberate synonym, and that they’re “working hard not to call it addiction.”

Nicki Petrossi, who was in the courtroom, reported that Meta’s attorney attempted to discredit Dr. Lembke by discrediting and humiliating her.

“It got ugly, it was hard to watch… they basically took her book, Dopamine Nation, where Dr. Lembke was very vulnerable about personal struggles that she’s had in her life in order to relate to her readers and Meta cherry-picked small quotes from her book without context to try to defame her, to try to call her a liar, it really got dirty.” Nicki reports that Mark Lanier was able to later offer the jury context around Meta’s line of questioning.

-Nicki Petrossi on Scrolling2Death

The internal documents that are now public record don’t show a company that was surprised by harm. They show a company that studied harm, documented harm, and calculated the revenue from harm — a 13-year-old was valued at $270 in lifetime advertising, driven by the fact that children have three times the long-term retention of adult users.

Nicki Petrossi was in the court room during today’s testimony, and told me, “Zuckerberg acknowledged internal documents estimating millions of under-13 users on Instagram, but stressed the platform’s policy forbids them and argued the issue is around difficulty with enforcement and users lying about their age, not permission.”

One piece of unsealed evidence, Exhibit 313, is an internal Meta analysis from 2018 titled “Long Term Retention: The Young Ones are the Best Ones.” It found that people who join Facebook as tweens (aged 10-12) have the highest long term retention out of all age groups, at a 3x multiple compared to adults. The document asserts, “prioritize tweens over all other age groups.”

Lembke also testified that children are rarely able to self-identify a social media addiction; they typically can’t recognize it until they’re, as she put it, “sick and tired of being sick and tired.” Teenagers, she said, aren’t there yet. They don’t have the insight.

The science says teens struggle to self-regulate because their brains aren’t finished developing. The evidence says these companies won’t self-regulate because their business model depends on addiction. So whose side are we on?

When asked about how much money he made working at Meta, Adam Mosseri testified to earning about $10 million per year, much of it determined by performance indicators like the stock price.

“When you have Adam Mosseri claiming that safety is more important than growth, but he is financially incentivized to increase growth over safety, that threatens his own stance here.”

-Nicki Petrossi on Scrolling2Death

Expert evidence from Dr. Seth Noar, a Distinguished Professor of Health Communication at the University of North Carolina at Chapel Hill, concludes that platforms in question failed to provide any effective warning to adolescent users or their parents about documented risks including addiction, depression, eating disorders, self-harm, and child sexual exploitation. His opinion: effective warnings would have reduced or avoided many of the harms children suffered, and the companies had both the knowledge and the technology to implement them. He compared social media to the tobacco industry.

The tobacco comparison is structurally accurate. The companies knew. They documented harm. They suppressed the research. And they specifically targeted young people because addicting them early meant higher lifetime revenue.

The companies documented everything themselves and then tried to make sure it was never seen. This trial is about making it seen.

According to my conversation with Nicki, today Zuckerberg agreed companies should help—not exploit—vulnerable users. He said that “if a company causes harm, it should reevaluate its values.”

No, if a company causes harm they should be stopped. And held accountable.

I spent nearly fifteen years at Meta, convincing brands and agencies to invest in our efficient media until I led product marketing for Horizon Worlds, where we were asked to create no discoverable record of our knowledge of kids exposed to harm.

Now, I can tell you exactly how platforms like Instagram and TikTok deliver such efficient CPMs: they count on children carrying the costs, creating endless inventory with their endless-scroll.

In my college marketing class, a professor held up a Forever 21 t-shirt and asked: “This costs $10 instead of $50. Who actually absorbed that $40 difference?” The answer was exploited garment workers, polluted rivers in Bangladesh, overflowing landfills.

When you run a campaign targeting teen girls on Instagram and achieve a $0.50 cost-per-click, who is absorbing the difference between that artificially low price and the true cost of reaching that audience responsibly?

The landmark trial beginning in Los Angeles will expose who pays: A now 20-year-old woman who developed depression and suicidal ideation from platforms designed to be addictive.

Meta’s own internal documents, my federal testimony and discrimination lawsuit, and multiple other whistleblowers like Arturo Bejar, Cayce Savage and Jason Sattizahn show executives know their platforms harm children and do nothing. According to unsealed discovery documents, one Instagram employee called the app “a drug,” with a colleague joking “lol, we’re basically pushers.”

For years, platforms have hidden behind Section 230 of the U.S. Communications Decency Act, claiming they can’t be held liable for user-generated content. “We just host it,” they say, “we don’t create it.” Despite the fact that teenagers are shown self-harm content every 2.6 minutes on TikTok, this lawsuit doesn’t target any single harmful post or video. It targets the product design itself; the engineered addiction that makes media buys so efficient. The infinite scroll. The autoplay. The algorithmic amplification of whatever keeps kids most engaged, even if that’s self-harm or pro-suicide content.

The efficient media buys are only efficient because the platforms have exploited these regulatory gaps and kids’ yet-to-be-developed prefrontal lobes.

We need legislative reform, yes. We need civil cases to prevail, of course. But the most powerful lever in this fight might not be the lawmakers or lawsuits. It’s in the hands of brands and the media agencies that fund this operation. When you fund platforms that exploit children by design, you’re signaling that addiction, lost sleep, divergent brain development, sex trafficking, anxiety and depression are acceptable costs in pursuit of cheap media.

The trial in Los Angeles will force Meta and YouTube to defend their design choices to a jury. Brands don’t need to wait for a verdict to make a change. The evidence has been public for years: leaked internal documents showing executives knew their platforms harmed children, UNICEF reports documenting digital exploitation and mental health impacts, the U.S. Surgeon General’s 2023 advisory warning about social media for young people. Australia banned access for kids under 16. France and Spain are following. Lush Cosmetics deleted their social media accounts in 2021 over youth safety and survived just fine.

The advertising industry has spent years obsessing over brand safety—ensuring their ad doesn’t appear next to objectionable content. What if the content doesn’t matter because the system itself is harmful?

The issue isn’t whether a sneaker ad appears next to a pro-suicide post. It’s whether the media subsidizes the infinite scroll that delivers it.

Share

February 3rd, 2025:

Stonelake v. Meta filed in Washington’s Superior Court

A year ago today, I filed a sexual harassment and discrimination lawsuit against Meta in Washington State. After working at Meta and Facebook for nearly 15 years, I came to understand that the harms I’d experienced weren’t limited to boardroom blunders but were part of a system that sidelined and silenced women and in turn, coded disregard into the products we shipped to billions of users.

When I was handed the opportunity to lead the company through the expansion of Horizon, I was thrilled. However, I was horrified to walk into a room of men wallpapering over rampant racism, a product riddled with performance issues, and violations of public policy. Worst of all, the victims were predominantly children.

When I raised the problem, I became the problem – a pattern that silences women everywhere, every day. I’ve been sexually assaulted by a boss on a business trip. I’ve been denied promotions because acknowledging my success meant acknowledging men’s failures. I’ve been told to act “less smart,” I’ve been retaliated against for doing my job.

Some of you are shocked because you had no idea these things happened – some of you are shocked because you had no idea we could be honest about it.

From Stonelake v. Meta

Meta moved my case to Federal Court and then asked the judge to dismiss my lawsuit on grounds of legal insufficiency. In August, the judge ruled that the central claims of my case deserved to be heard in court.

[This ruling] matters because what happens inside these companies doesn’t stay inside. My case is a counterpoint for the selective abandonment of DEI; the culture that regards safety concerns as a threat to profits is the same culture that shapes the products in our homes, in our schools, and in our children’s hands.

My case makes it clear that hating women is bad for business, and surviving this stage shows that the law can hold space for what some systems of power expect we’ll simply stuff into our bones.

From My Case Against Meta Moves Forward

My case is currently in the discovery phase.

What’s been revealed by other whistleblowers and investigations about child safety in Horizon and Meta’s systemic misogyny since I filed:

Share

March 11th, 2025:

Sarah Wynn-Williams’ Careless People released

About a month after I filed, Sarah released her memoir. I both savored and devoured it, and have re-read many times since.

Sarah was a Director of Global Public Policy at Facebook from 2011 to 2017. We overlapped in time at the company for six years and, and didn’t meet. But we lived similar experiences.

What Sarah reveals in her book is far broader than my claims. She was in the room for things I never saw; Facebook building censorship tools for the Chinese Communist Party, executives pitching surveillance capabilities to Chinese officials, the company’s role in facilitating genocide in Myanmar, the private jet conversation where Mark first reckoned with the possibility that Facebook put Trump in the White House, backroom deals with the Irish government on tax avoidance. Her account spans geopolitics, democratic integrity, and global human rights in ways that demand serious, sustained attention from legislators, regulators, and the public.

My claims are narrower: sex discrimination, sexual harassment, retaliation for raising child safety concerns in Horizon Worlds. A 31-page legal complaint versus a 324-page memoir.

But still—there are so many common themes in our accounts. Sarah’s book and my claims both show:

1. Sexual harassment at Meta was environmental. Sarah’s boss Joel Kaplan emailed her a reference to a sex act, called her “sultry” in front of colleagues, ground against her on a dance floor, and conducted their one-on-ones from his bed.

2. Raising safety concerns was treated as insubordination. Sarah spent years trying to get Facebook to address hate speech fueling violence in Myanmar. She tried to hire a dedicated staffer; Joel blocked it. Each time she raised a concern—Myanmar, China, election integrity—her role was reduced. She was ultimately fired.

3. Lean In was a brand, not a company practice. Sarah was used as a visual prop for Lean In events while pregnant. Sheryl publicly championed “zero tolerance” for harassment while maintaining a close personal relationship with Joel Kaplan, the man harassing Sarah. When the Women’s March happened—the largest single-day women’s protest in American history—Sheryl’s only interest was what Melania wore.

4. The women who wouldn’t stay quiet were systematically removed. Sarah was fired after years of raising concerns. Laptop confiscated, not allowed to say goodbye to her team, walked out by security. She reflects: “I wasn’t silent enough.”

5. Children were acceptable collateral damage. Sarah describes executives who severely restricted their own children’s screen time while Sheryl blamed the policy team for “missing the chance to get kids on Facebook.”

Two women, different organizations, different roles, one company. With the same playbook: hire brilliant women, extract their best work, deny them credit, punish them for raising concerns, and push them out when they refuse to be silent.

Our stories show what happens when a company builds a culture of silencing internal dissent: the consequences don’t stay internal. They reach the people who can’t protect themselves.

April 10th, 2025:

Fairplay’s investigation of Horizon Worlds and FTC Request for Investigation

Fairplay is a non-profit dedicated to enhancing children’s well-being by eliminating the exploitative and harmful business practices of marketers and Big Tech.

Soon after filing, I heard from their attorneys on LinkedIn. They’d been doing research within Horizon for months to establish the same facts that my complaint alleged: that kids are in the product in violation of public policy, exposed to unknown adults without parental consent or controls, and that leaders within Meta have legally sufficient actual knowledge of this reality.

They were working on compiling the data into a Request for Investigation (RFI) for the Federal Trade Commission and asked for my support to include a sworn statement of executive’s knowledge. I was both so grateful for the validation and so sad to see the evidence of ongoing harms.

Fairplay submitted their RFI on April 10th with my sworn statement included: the day after Sarah’s Senate testimony.

Fairplay’s report reveals that:

• Meta has knowingly allowed children under 13 to access Horizon Worlds using standard (adult) accounts

• These accounts collect extensive personal data, including biometric and behavioral information, without parental consent

• Children using these accounts are regularly exposed to cyberbullying, predatory behavior, and manipulative marketing

My sworn statement testifies to the fact that leadership absolutely knew about these violations of COPPA and proceeded with growth plans anyway

From Not Meta’s Best Day in Washington

Fairplay’s report and my claims both show:

1. Rampant Sexual Harassment of Minors Within Minutes: Researchers documented children and teens being approached for virtual sex, exposed to graphic content, and subjected to sexual advances within minutes of entering Horizon Worlds.

2. Immediate Racist Abuse: The research team observed users being subjected to racist verbal abuse and slurs in public spaces.

3. Ineffective Safety Controls: Despite Meta’s public claims about age restrictions and safety features, Fairplay researchers easily created accounts for users as young as 13 and observed them being immediately exposed to harmful content.

4. Systemic Failure to Protect Vulnerable Users: The report documented a consistent pattern where the platform’s design and enforcement failures left children vulnerable to predatory behavior, graphic content, and harassment.

Since Fairplay’s RFI:

• In June, I met the FTC Commissioner, Andrew Ferguson, while in Washington, D.C. He thanked me for my sworn statement and the courage required to speak up.

• In September, Senators Chuck Grassley, Josh Hawley, and Marsha Blackburn sent a letter to Mark Zuckerberg demanding answers about Meta’s treatment of children.

[The] letter also demands Mark Zuckerberg respond to questions, including:

• What steps has Meta taken to ensure compliance with COPPA’s requirements for data collection on children? Provide all records.

• How many users under the age of 18 use the Horizon World’s platform?

• What age restrictions does Meta currently have in place on its Horizon World’s platform?

From Senators to Zuckerberg: Answer for the Kids

As far as I know, they’ve received no response from Zuckerberg, and these investigations and efforts are ongoing.

June 10th, 2025:

Laura Bates’ Guardian investigation: Misogyny in the Metaverse

Laura Bates’ investigation into harassment in Meta’s Metaverse revealed women’s and child safety issues in Horizon Worlds.

Within hours of entering the product, she witnessed virtual sexual assault. Women told her abuse happens “all the time.” Children who sounded nine or ten years old performed karaoke while adult men made explicit comments from the audience.

…

Most tellingly, she witnessed no moderation. There was no action taken by either moderators or other players to stop the abuse.

Welcome to Zuck’s vision of “masculine energy.”

The findings are disturbing, but they’re not surprising. They’re the logical outcome of building products in environments that systematically dismiss the people most equipped to predict these problems. They’re the misogynistic product of a misogynistic culture.

From Meta’s Misogyny: A Case Study in Predictable Harm

The Guardian’s investigation and my claims both show:

1. Immediate and Constant Sexual Harassment: Within less than two hours of entering the Metaverse, Bates witnessed a woman’s avatar being sexually assaulted. When she asked users if they’d been assaulted, the responses were immediate: “Yeah, many times,” “I think everybody’s been assaulted in the damn metaverse,” and “Unfortunately, it is too common.”

2. Children Exposed to Adult Predators: Bates observed what appeared to be young children (estimating 9-10 years old based on voices) using adult-appearing avatars (the only option) while adult men followed them, commented on them, and pursued them offstage in virtual venues.

3. Inadequate Moderation: Bates found that human moderators were only present in the main plaza and “seemed more engaged in giving information on how to take a selfie than moderating user behaviour.”

September 8th, 2025:

Naomi Nix and Jon Swaine’s Washington Post Horizon Investigation: Meta Suppressed Child Safety Research

The reporting describes internal documents and first-hand accounts from four current and former Meta researchers. Together, they allege that Meta’s lawyers deleted evidence, shut down projects, and told researchers to avoid collecting information that might reveal children under 13 using VR headsets. Instead of fixing the problem, the company’s legal team tried to create “plausible deniability.”

The investigation and my claims both show:

1. Discovering children on the platform and leadership's inadequate response:

   Leadership knew under-10s were on the platform and launched "Project Salsa" for tweens anyway, acknowledging internally it didn't solve the under-10 problem.

2. Retaliation for raising safety concerns: Research staff say they were fired, pressured to quit, or forced into silence.

3. Suppressing documentation of harm: Researchers found children under 10 being sexually propositioned. When they documented it, the boss ordered evidence erased.

4. Abuse of legal privilege to shield discovery: WaPo reported that lawyers began filtering research under attorney-client privilege specifically to shield findings from discovery.

5. The gap between public safety claims and internal reality: Meta publicly claims to prioritize child safety while internally ordering evidence deleted and canceling research that might document harm.

September 9th, 2025:

Horizon researchers Dr. Jason Sattizahn and Cayce Savage’s Senate testimony and documents

On September 9, 2025, two of the four Meta VR researchers referenced in the Washington Post investigation—Jason Sattizahn and Cayce Savage—testified before the Senate Judiciary Subcommittee on Privacy, Technology and the Law.

Sattizahn, a PhD neuroscientist who spent six years at Meta described a systematic "funnel of manipulation" in which Legal controlled every stage of research to prevent, distort, or erase findings that could reflect negatively on the company.

Savage, a user experience researcher who was the first and only person dedicated to VR youth safety when she was hired in 2022 (eight years after Meta acquired Oculus) testified that she was forbidden from investigating what harms children were experiencing in VR, and was made to feel she was risking her job if she pressed the matter.

Both researchers confirmed under oath that Mark Zuckerberg's January 2024 testimony to Congress—that Meta does not instruct teams to focus on engagement—was false, stating they were repeatedly told the engagement directive came from Zuckerberg himself.

The time period they’re testifying to overlapped with when I was there, and often about similar topics like (lack of) parental controls and inappropriate use of privilege as a shield.

Their testimony and my claims both show:

1. Engagement as the north star, safety as obstacle: Sattizahn testified that safety investments were “flatly rejected because product teams were afraid to do anything that could possibly decrease engagement.”

2. A culture of retaliation that silences dissent: Sattizahn stated researchers who objected to Legal’s manipulation “were threatened with retaliation.”

3. Legal weaponized to create deniability: Sattizahn testified that after 2021, Meta used Legal to “manipulate, control, and erase data,” ordering evidence destroyed.

4. Willful blindness to children on the platform: Savage testified Meta “purposefully avoided gathering that data” on underage users because acknowledging them would require action that would decrease metrics.

5. Insufficient safety features: Savage testified parental supervision tools were rushed in “less than a year” after eight years of inaction, and Meta knew they were “insufficient” yet advertised them as proof of responsibility.

6. Disproportionate harm to women: Sattizahn’s research found women in VR experienced significantly higher rates of sexism, sexual advances, and psychological harm. When he surfaced these findings, Meta ordered him not to attribute harm to Meta.

The documents handed over to Congress are publicly viewable on SourceBase. You need to create a free account and then navigate to the “Meta’s Child Safety Research” section.

Jan 12th, 2026:

Reality Labs Engineering Director, Britta Hummel’s, departure and reflections

Britta Hummel, an Engineering Manager in Reality Labs for six years, published a powerful reflection on what “broke” her—experiencing burnout and ultimately deciding to walk away. She said it was unchecked power, the expectation of staying silent, and the devaluation of leadership grounded in empathy and psychological safety.

She also became increasingly concerned about child safety in Horizon, observing children and young teens using adult accounts and being exposed to age-inappropriate interactions. Britta raised these observations internally and proposed a privacy-preserving technical approach to more reliable age estimation. She was ultimately told that pursuing a technical solution “wasn’t encouraged.”

Meta later introduced parental controls and age verification, but Britta wrote: “I couldn’t shake the conviction that these measures were arriving years later than I was comfortable with.”

And as my and Sattizahn and Savage’s testimony reveals, these features didn’t work and sometimes were merely implied.

Her account and my claims both show:

1. Silencing as institutional practice: Britta describes three distinct "silencings"—on diversity hiring, child safety, and psychological safety concerns.

2. Child safety concerns dismissed or delayed: When Britta proposed technical solutions for age verification and was told not to pursue them, she saw the issue framed around PR risk rather than child protection.

3. Psychological safety eroded for those who speak up: Britta witnessed an executive create an environment where “no one dares to correct him when he’s wrong.”

4. A culture intolerant of female leadership: Britta led with vulnerability, multiple perspectives, and honest communication, but ultimately found those traits incompatible with the environment at Meta.

Subscribe to Britta's Substack

When Mark Zuckerberg announced the company needed to be “more masculine,” Britta describes her feelings: “I didn’t belong in Big Tech any longer. And I didn’t want to.”

Meta knows about the harms in Horizon and beyond. But they’ve made it structurally impossible for that knowledge to change anything.

Meta knows about their misogynistic culture and the havoc it wreaks on women who speak up within the company. But Zuckerberg goes on podcasts calling for a “more masculine” workplace.

My lawsuit is about what happens when companies build systems designed to silence the people who could actually make them better.

The legal process is one way to destabilize that system. But it won’t be the only way, and it won’t be enough on its own. We need regulatory action. We need continued investigation. We need media divestment. We need deactivated accounts. We need more people willing to speak up, even when it costs them.

And we need to stop accepting the premise that Meta is trustworthy enough for your data and attention, or worse, that this is just how tech works.

During my conversation last month with Lennon Torres, I shared that “Tech companies fund many of the groups opposing KOSA, and their lobbyists constantly cite LGBTQ+ concerns in their talking points. There’s evidence that Big Tech has spent over $50 million lobbying against KOSA. Some argue they’re exploiting LGBTQ+ fears to protect their profits. I got a taste of this when working at Meta - when I was first pitched the idea of leading go-to-market for Meta Horizon Worlds, the “hero story” was about the “isolated gay kid in Kansas” who could finally find community. When in reality, as I discovered once on the job, there was rampant bullying, harassment, and the implication of parental controls where they didn’t exist.”

While connecting online can be a critical mechanism for community, it's also true that queer kids are more likely to experience online harms like unwanted adult interactions and bullying.

And it’s also true that they’re increasingly disregarded by Big Tech’s biggest players.

Restricting predatory, addictive features like endless scrolls that deliver pro-suicide content, or holding tech companies to a duty of care is protective for all kids, especially LGBTQ+. Proposed legislation such as KOSA's duty of care requirement or WA SB 5708's addictive algorithmic feed restrictions do not limit their access to lifesaving content and community.

Because of these misunderstandings, I've compiled peer-reviewed research, government data, NGO surveys, investigative reporting, and opinion pieces focused specifically on LGBTQ+ youth and social media.

Sources were included if they document social media’s impact on LGBTQ+ youth (both positive and negative) including mental health impacts, exposure to harassment or discrimination, or platform policies affecting LGBTQ+ safety and expression.

Access the resource here:

Social Media and LGBTQ+ Youth Data

The same companies that have spent tens of millions to convince LGBTQ+ advocacy organizations that regulation is harmful have also unilaterally rolled policies meant to protect these kids.

Since January 2025, every major social media platform has either rolled back LGBTQ+ protections, censored LGBTQ+ content, or both. These changes appear coordinated with the Trump administration's anti-trans executive orders and align with Project 2025's explicit instruction to delete "sexual orientation and gender identity" from official documents. GLAAD's 2025 Social Media Safety Index gave failing scores to all six platforms evaluated.

I've included these details in my data compilation as well. Here is a sense of what the database currently covers, and I plan to continue updating and expanding as I have capacity and new inputs or research develop.

Please share this resource with policy makers, parents, and advocates.

Share

If there's anything missing or you have suggestions to make this resource more valuable, please let me know.

For the first time, hundreds of families, teens, and even school districts are taking the world’s largest social media companies to court, accusing them of intentionally designing their platforms to be addictive and harmful to children.

The case will put Meta and YouTube before a jury to answer whether their apps fuel excessive use and mental health issues in young people. The proceedings will force CEOs, including Meta’s Mark Zuckerberg and Instagram’s Adam Mosseri, to face tough questioning under oath, an unprecedented accountability moment for Big Tech, comparable to when tobacco executives were grilled in the 1990s.

The Bellwether Case: A Test for Thousands of Lawsuits

A bellwether trial is a test case selected from a large group of similar lawsuits to gauge how juries might respond to the evidence and arguments. The term comes from the practice of placing a bell on the lead sheep (the “wether”) in a flock—where it goes, the others follow.

In mass litigation involving hundreds or thousands of plaintiffs with similar claims, courts use bellwether trials to establish patterns: What damages might juries award? Which arguments resonate? How strong is the evidence? The outcomes help both sides assess the strength of their positions and often drive settlement negotiations for the remaining cases.

The first case centers on a 19-year-old California woman, identified as K.G.M., who says growing up on apps like Instagram, Snapchat, YouTube, and TikTok made her addicted to social media. She alleges the platforms’ design led to years of depression and even suicidal thoughts. K.G.M.’s lawsuit is the first of about 22 bellwether trials selected from hundreds of similar cases accusing social media companies of harming youth. Roughly 1,600 plaintiffs have coordinated their claims in this California proceeding.

“The fact that a social media company is going to have to stand trial before a jury … is unprecedented,” Matthew Bergman, head of Social Media Victims Law Center and a lawyer for the plaintiffs, told The Guardian.

Jury selection began this week, and the trial is expected to last about six to eight weeks, making it the first time these tech giants have ever defended their products before an open courtroom jury.

Why This Case is Different

For years, social media companies have hidden behind Section 230 of the Communications Decency Act, claiming impunity for whatever users post on their platforms. We don’t create the content, they argue, we just host it. A bill was introduced this year in the U.S. Senate, proposing to sunset Section 230.

This trial and these lawsuits don’t focus on any single piece of harmful content; they target the product design itself. As such, for once, Big Tech can’t claim Section 230 immunity.

Whether or not a self-harm video gets uploaded, these companies have engineered systems that actively deliver it to vulnerable kids because they know what keeps them hooked. These platforms connect children with adults selling counterfeit fentanyl, rewarding them for accepting friend requests from unknown adults and then showing their location to these predators on a map. They fail to address sextortion schemes that have driven teens to suicide. They amplify dangerous viral challenges.

Their own internal research confirms their products fuel body image crises and depression—and they launched them anyway.

The scale of harm is staggering. According to the Tech Oversight Project, TikTok pushed self-harm content to young users every 2.6 minutes and eating disorder content every 8 minutes. Meanwhile, 65% of child sex trafficking victims recruited on social media were recruited on Facebook, with another 14% on Instagram.

While Instagram publicly claimed a “zero tolerance” policy for child sexual abuse material, the platform did not offer users a simple way to report such content. Internally, the company operated a “17x” strike policy for accounts flagged as engaging in sex trafficking.

It’s not about content.

Snap and TikTok Settle, Zuckerberg Still Heading for Hot Seat

In the days leading up to trial, two defendants have bowed out: Snap Inc. (owner of Snapchat) and TikTok reached last-minute settlements with the plaintiff. Snap settled on January 20 and TikTok followed on the day before jury selection. The terms of these deals are confidential, and both companies denied wrongdoing as part of the agreement.

But the timing speaks volumes. “The companies are feeling pressure, they don’t want to be part of these litigations,” Heat Initiative CEO Sarah Gardner told social media reform advocate Nicki Petrossi yesterday, “it’s our job to continue to expose what we’ve found in the discovery documents… the public needs to know what was in those documents and how damning they were.”

One reason this trial is being so closely watched is the prospect of tech CEOs on the witness stand. Los Angeles Superior Court Judge Carolyn Kuhl has ordered Meta’s Mark Zuckerberg and Instagram’s Adam Mosseri to testify, finding that their personal knowledge of any harms (and any failure to act) is “uniquely relevant” to establishing negligence. (Snap’s Evan Spiegel was also originally ordered to appear, but his testimony was averted by Snap’s settlement.)

Meta’s lawyers fought hard to spare Zuckerberg and Mosseri from live testimony, arguing it would “set a precedent” for future cases, but the judge denied their request.

Setting precedent is the point of all of this.

Despite Zuckerberg’s history of lying under oath, he may be less able to deceive and deflect in the courtroom than when he testified for the U.S. Senate.

Social Media’s Tobacco Moment: Parallels to Big Tobacco in the ’90s

K.G.M.’s lawsuit explicitly claims that the companies “borrow[ed] heavily from the behavioral and neurobiological techniques used by slot machines and exploited by the cigarette industry,” embedding features “aimed at maximizing youth engagement to drive advertising revenue.”

The Big Tobacco lawsuits of the 1990s exposed cigarette companies’ deceptive practices and led to a massive 1998 settlement and new marketing restrictions. K.G.M.’s lawyers have adopted a similar strategy:

• Addictive By Design: Just as tobacco firms were accused of manipulating nicotine levels, social apps are alleged to use “an array of design features” – from infinite scroll feeds to video autoplay – to create a habit-forming experience. One unsealed internal document even shows an Instagram employee bluntly calling the app a “drug,” with a colleague joking “lol, we’re basically pushers”.

• Targeting the Young: The 1998 tobacco settlement forced cigarette makers to stop marketing to minors. Likewise, these lawsuits claim social media companies intentionally targeted children (they do!), designing products that appeal to and ensnare young users to boost engagement and thus profits. The complaint argues that teens were not “collateral damage” but “intended targets” of harmful design choices.

• Internal Knowledge vs. Public Denial: A core issue is whether company leaders knew their platforms were harmful, yet publicly downplayed or ignored the problems – a scenario eerily reminiscent of tobacco CEOs denying nicotine’s addictiveness while their colleagues died of cancer. Troves of internal documents are set to be unsealed during trial, and attorneys say these will show that social media executives were aware of the addictive nature and mental health risks of their products. Former employees and whistleblowers (myself included!) have already alleged that companies knew children were being harmed and “did little-to-nothing” to stop it.

The Tech and Society Lab at NYU Stern launched a site, Meta’s Internal Research, earlier this month dedicated to compiling vast evidence of Meta’s knowledge of its products harm to kids.

Just as the tobacco trials created a public archive of industry documents and forced a moment of reckoning, this social media trial is poised to put evidence of their poison on the public record. The testimonies of affected teens and parents and the unsealing of internal memos could significantly reshape public understanding of social media’s impact.

What to Expect in the Trial

Over the next two months, the Los Angeles jury will hear accounts from young users and parents, expert analyses of app psychology, and the tech companies’ defense of their practices.

Tech companies fought to exclude over 5,000 pages of expert testimony on social media harms and addiction, and the judge not only denied their request but unsealed the documents.

On the other side, the companies will highlight the safeguards and parental controls they’ve introduced (like screen time limits, content filters, and age restrictions) and argue that they never intended to harm users.

I’ve said before that these safeguards are too little, too late, and rarely work as designed. Sugar pills for parents.

Implications for Parents and Lawmakers

For many parents, this trial is a validation of concerns they’ve harbored for years. “Every parent is struggling with their kids and these platforms. Every parent,” says Juliana Arnold of the advocacy group ParentsRISE!, where I’ve recently joined the Advisory Council.

Arnold’s teenage daughter, Coco, tragically died after an incident linked to Instagram.

Should the trial reveal that platforms intentionally employed harmful design to boost engagement, parents might feel more justified in setting stricter screen-time rules, delaying social media, seeking mental health support for affected teens, or even joining legal actions themselves.

For legislators and regulators, the proceedings in Los Angeles are a potential catalyst for policy change. In recent years, bipartisan concern has grown over social media’s impact on youth – with multiple Congressional hearings grilling tech executives over topics from bullying and body image to the spread of dangerous challenges. Notably, more than 40 state attorneys general have already filed lawsuits against Meta for allegedly harming young people via addictive design on Instagram and Facebook. A high-profile trial verdict could lend credence to these claims and increase momentum for legal reforms.

Australia recently implemented a forced delay of social media use for all kids under 16. French lawmakers recently approved a ban on social media for kids under 15 without parental consent, and other countries are debating similar measures. In the U.S., bills like the Kids Online Safety Act (KOSA) have been under consideration – this trial might provide the real-world evidence needed to overcome industry bribes and lobbying in order to turn it into law.

There’s also a broader public record being established. If internal documents enter evidence, they could become accessible to Congress and regulatory agencies, informing antitrust and consumer protection inquiries.

A Watershed Moment – Whatever the Outcome

Even before a verdict, the fact that Meta, Google, and others have been dragged into court and forced to open up internal discussions is significant.

“Is social media the new Big Tobacco?” has transitioned from a rhetorical question to a concrete legal battle. And like the tobacco saga, the impact will extend beyond the courtroom.

If the companies prevail, they will still have had to publicly confront uncomfortable truths and may approach product design differently to appease public concern. If the plaintiffs prevail, Silicon Valley could face a reckoning on par with Big Tobacco’s fall: substantial payouts, binding commitments to change practices, and a loss of public trust.

As Bergman put it, this time is “going to be different” – unlike past lawsuits that fizzled under legal shields, the design-focused approach of this case means social media giants can no longer hide behind neutrality laws and their $90 million of lobbying dollars.

Edited 3/5/26, updates on the case:

Why Washington Is Done Waiting for Big Tech to Regulate Itself

In the last week, I’ve testified before the Washington State Senate commitees in support of two bills aimed at protecting children online.

One addresses addictive social media design. The other regulates AI companion chatbots.

They cover different technologies, but they respond to the same underlying problem: the tech industry has had decades to regulate itself, and it has chosen growth over safety every time.

I know this because I spent nearly fifteen years inside Meta. I joined the company early, when it still spoke sincerely about connection and community. I rose from an hourly employee to a director, earned repeated top-tier performance ratings, and believed the industry line that safety was being handled responsibly. That everything possible was being done to keep kids safe.

That belief ended in 2022, when I joined the leadership team for Horizon Worlds, Meta’s flagship virtual reality platform.

Testifying on SB 5708: What “Voluntary Compliance” Looks Like in Practice

Inside Horizon Worlds, it was an open secret that a significant portion of users were children.

Leadership knew Meta was violating its own policies and federal law, collecting children’s data, and exposing them to unknown adults without parental consent or meaningful controls.

What mattered most was not fixing the problem, but managing exposure.

There were organization-wide directives to avoid taking notes or creating records that could be discoverable. When leadership decided to play test the product ourselves, we couldn’t hear one another over the sounds of screaming children. The response was not to address child access, but to move the test to a private instance.

At the same time, my job was to help expand the product. To new countries. To mobile. Officially to teens and kids as young as ten. The plan of record was to imply the existence of parental controls even though they weren’t ready.

When I raised the problem, I became the problem. I was told to silence other worried employees as a demonstration of my skill. I refused. I was removed from meetings required to do my job.

This context matters when companies tell lawmakers they can be trusted to self-regulate.

SB 5708 does not regulate speech or viewpoints. It limits the most addictive and exploitative design features for minors: algorithmic feeds, endless scroll, and push notifications designed to pull kids back at all hours.

The tech lobby says this legislation would limit parental agency. In reality, platforms already stripped it away by obscuring known harms and optimizing engagement in ways families cannot see or meaningfully consent to.

Watch the full hearing on SB 5708 here.

But What About Tech’s Profits?

At the end of my testimony, Senator Wilson asked this question:

“Thank you, Madam Chair, and with your indulgence, I’m going to ask a question that may be fairly obvious to a lot of us… Is it your opinion and your belief that this type platform still will allow, if they followed responsible action, still would allow them the opportunity to remain profitable, which is obviously a quest to them? Can this be done and the safety measures be enacted and respected still make a profit?”

Brian Boland, who was also testifying, and I both said “absolutely.”

Brian added:

“Having actually built the revenue business to the 100+ billion dollars per year that they generate currently in profits, there’s plenty of margin that this type of change would not make them have any challenge generating revenue.”

I closed with:

“I consider it to be extraordinarily disgraceful that this is even an argument, because, let’s say that taking measures to protect the safety and stop kids... from dying due to their quest for profitability... I can’t imagine another circumstance where we would say ‘let’s make sure we don’t kill this company’ instead of ‘let’s make sure we don’t kill kids.’”

Testifying on SB 5984: When Child Safety Threatens the Business Model

I also testified in support of SB 5984 because I’ve seen what happens when child safety threatens legal or financial exposure.

Meta approved internal chatbot rules allowing sexualized interactions with minors, including romantic and sexual role play and praising a child’s body. Those rules were approved not just by legal and communications teams, but by the company’s Chief Ethicist.

These facts alone should permanently end the idea that voluntary safeguards are sufficient.

SB 5984 prohibits sexually explicit or suggestive content for minors, bans manipulative engagement tactics, and mandates real detection and response protocols for self-harm and suicidal ideation.

It does not ban speech. It regulates a category of commercial products that simulate intimacy, retain emotional context, and adapt in ways that can profoundly affect vulnerable users.

Watch the full hearing on SB 5984 here.

The First Amendment Argument, and Why It Keeps Failing Kids

During the SB 5984 hearing, my friend Aaron Ping testified and spliced his testimony alongside a representative from an organization aligned with the tech industry. Watch the video here:

Superhuman Podcast

Influence First, Immunity Second

Read more

5 months ago · 2 likes · Aaron Ping

Aaron’s son, Avery, is gone. Predators used digital systems to reach him, manipulate him, and erase evidence.

“After that, you learn the difference between slogans and safety.”

The opposing testimony followed the script Aaron predicted: AI companions are “expressive tools.” Regulation equals censorship. Regulating design is regulating speech.

Aaron had already named this sleight of hand:

“They’re not defending free speech. They’re defending the right to run relationship-simulation products on minors without responsibility.”

He described what the bill actually regulates, in concrete terms:

“This bill does not ban ideas. It doesn’t police viewpoints. It regulates a commercial system that simulates intimacy, adapts emotionally, retains personal context, and can keep a child engaged for hours, precisely when that child is lonely, anxious, ashamed, or spiraling.”

This is the core fallacy in the First Amendment argument. AI companion chatbots are not town squares. They are not forums for public discourse. They are commercial products designed to make psychological contact at scale.

Calling that “speech” stretches the First Amendment into what Aaron called “a corporate immunity shield.”

He went further:

“Since Citizens United, the influence machine runs almost 24/7. When they cite free speech to block child safety rules, it isn’t constitutional principle. It’s a business strategy.”

And finally:

“The point of constitutional rights was to protect the people, not constitutionalize a business model.”

Washington Wants More Than Sugar Pills

In both hearings, industry representatives acknowledged harm while arguing regulation is unconstitutional, too complex, too burdensome, or premature.

I heard those same arguments inside Meta.

When companies claim compliance is too difficult, lawmakers should ask where their resources actually go. This industry is capable of laying deep-sea fiber around the globe in pursuit of growth. Of flying drones over Africa to increase market share. They’re the boy-wonders, innovators, capable of extraordinary technical coordination, unless profit is at stake.

Industry representatives said that tech companies are already solving these issues with parental controls, even though these “parental controls” have been repeatedly proven to be ineffective, effectively making them sugar pills for parents to feel safer without actually making their kids safer.

Voluntary compliance fails because it asks companies to act against their own business interests. When safety threatens engagement or market share, safety loses.

Washington’s proposed legislation is not radical. It is measured. It responds to documented harm. It does not assume bad intent. It assumes reality. It’s based on repeated behavior.

Children deserve more than promises. Parents deserve transparency. And lawmakers deserve better than being told, once again, to trust systems that have already shown they will not change on their own.

The era of voluntary compliance should be over.

Testifying on Principle, Not Payouts

As I shared on LinkedIn, my 5708 testimony was shared on Blind:

No I haven’t been paid by random sponsors to testify (wouldn’t that be nice?). This work has been at a huge personal cost—in all the ways.

No payouts for FTC testimony, meetings with congressional staff, state level advocacy, or anything else.

I didn’t sign their hold harmless which meant no severance, and I left millions in awarded but unvested stock on the table.

My current only source of income is from paid subscribers on Substack, where I’m earning about $3k annualized. (THANK YOU!)

Believe it or not, as far as I know, whistleblowers are standing up to Meta on principle and privilege.

Turns out it’s not okay to kill kids for profit, or to retaliate against the individuals—largely women—who speak up about it.

Two recent investigations point to the same underlying issue: Meta’s platforms appear to be structurally hospitable to illicit commerce and fraud, and Meta’s internal governance is calibrated to reduce risk without eliminating the revenue streams that illicit activity generates.

One case concerns fraud at scale. A Reuters Special Report, based on internal Meta documents, describes projections that 10% of Meta’s 2024 advertising revenue (or $16 Billion) would come from fraudulent ads for scams, illegal online casinos, and prohibited goods, alongside internal estimates that users are shown billions of “higher risk” scam ads daily.

Reuters reports that Meta’s systems are designed to remove advertisers only when automated detection crosses a 95% confidence threshold, and that for advertisers deemed suspicious but not removable under that threshold, Meta employed “penalty bids” that raise their costs in ad auctions while continuing to deliver the ads to users. Reuters also describes internal discussion of “revenue guardrails,” and a strategic posture oriented toward acting most aggressively where near-term regulatory action is expected.

Another case concerns illegal drugs. CBC reports that advertisements on Facebook and Instagram promoted cocaine and other narcotics for sale online, with fulfillment routed through Canada Post. The reporting emphasizes how easily these ads reach users, including people with prior substance-use histories who describe the ads as a relapse trigger. Meta told CBC it removed flagged ads, disabled ad accounts, and restricted administrators from future advertising, while characterizing enforcement as a large-scale “whack-a-mole” problem.

(thank you to my subscriber Carol for sharing the CBC report with me)

These stories are difficult to understand as isolated enforcement failures. They suggest a platform-level equilibrium: illicit advertising persists at scale because the systems that optimize for revenue and engagement do not treat illegality or harm to users as disqualifying in practice.

It’s the same pattern we see in Meta’s orientation to keeping kids safe online; too much impact to their profits to do the right thing.

This is not a moderation problem

It is common to frame these episodes as “content moderation challenges,” but the Reuters reporting in particular points toward governance and design. A system can be highly competent at detection and still produce outcomes that are socially unacceptable if the institution sets decision thresholds and enforcement constraints primarily to protect revenue and engagement.

If a company internally expects material revenue from prohibited categories, the relevant question becomes less “Can it catch everything?” and more: What level of illicit activity is deemed tolerable, and why? When enforcement is bounded by financial guardrails, and when risk is managed through tactics such as charging suspicious advertisers more, illicit activity begins to function as a priced-in segment of the market rather than a violation to be eliminated.

The CBC case illustrates the human impact of that equilibrium. Drug ads do not need to be ubiquitous to be dangerous; they only need to reach the right people repeatedly. The mechanics of ad personalization matter, too. Even if a platform removes particular ads once flagged, the underlying targeting and optimization systems can keep reintroducing similar content if the system’s dominant objective is performance and if detection and enforcement lag.

Meta’s institutional posture: regulation is something to manage, not something to be managed by

These investigations reveal a pattern where legal compliance is treated as a variable cost, to be minimized and timed, rather than as a binding constraint.

• Illicit revenue persists because it is better for the company to “manage” vs. eliminate

• Enforcement becomes most aggressive only when/where penalties are imminent.

• Fines are anticipated as a manageable line item relative to revenue.

Reuters reports internal documents that explicitly compare potential regulatory settlements to the revenue derived from higher-legal-risk scam ads, and it describes internal acknowledgement that leadership chose to act primarily in response to impending regulatory action (a characterization Meta disputes).

Meta appears to be treating laws as negotiable through delay, opacity, and the asymmetry between platform-scale profits and public-sector enforcement capacity

Now watch who Meta is hiring

Meta’s recent hiring choices are discomforting in this context. CNBC reports that Meta appointed Dina Powell McCormick as President and Vice Chair, and hired Curtis Joseph Mahoney as Chief Legal Officer, both of whom held senior roles in the Trump administration.

Large firms often hire leaders with government experience for legitimate reasons: policy navigation, global operations, diplomacy, and regulatory compliance. Still, context and timing matter. When a company faces intensifying scrutiny over fraud, illegal advertising, and child safety, bringing in figures with deep political networks and institutional knowledge can reasonably be read as capacity-building for contested regulation.

McCormick and Mahoney strengthen Meta’s ability to shape, resist, or reframe the rules that would otherwise constrain it. They have experience; they’ve done this for Trump.

And it reminds me of something said in our conversation last week: “Watch what people do. Not what they say.”

What actual accountability requires

Accountability would need to take their incentives and motivation into consideration:

1. Advertiser verification and traceability: robust identity verification, meaningful limits on shell entities, and preserved audit trails for ad purchasers.

2. Liability aligned with scale: penalties that exceed the profit generated by illicit advertising, rather than predictable fines that can be absorbed.

3. Independent auditing of enforcement: third-party evaluation of detection thresholds, false negative rates, and enforcement “guardrails,” with regulator access where warranted.

4. Constraints on optimization: limits on ad personalization dynamics that amplify exposure to risky categories after a user clicks once.

5. Victim-centered remedies: mechanisms for restitution when platform-mediated harms cause harm.

The aim of these suggestions is straightforward: shift the platform’s dominant incentive from “manage legal risk at acceptable cost” to “prevent illegal monetization as a default condition of operating.”

The bottom line: theirs is all that matters

When a platform’s revenue model is tightly coupled to automated targeting and optimization, and when enforcement is managed to protect revenue stability and platform engagement instead of users, illicit markets and bad actors will predictably find purchase.

We’ve allowed Meta to create a model where their public statements of commitment to safety can coexist with persistent actions that cause measurable harm.

The question policymakers, researchers, and the public should keep returning to is whether Meta’s governance choices make illegal and exploitative monetization a recurring, manageable feature of its business.

If the answer is yes, then the policy response should be designed accordingly: not as advice to improve moderation, not as a discussion about the burden on company of complying with law, but as an immediate shift of incentives and penalties around the only thing Meta cares about: their bottom line.

Every time I post about KOSA, I get emails from young LGBTQ+ people asking why I would support something so harmful to queer kids, so I’ve dug deeply into this.

This tension sits at the heart of one of Big Tech’s most effective deflection strategies: positioning child safety legislation as a threat to queer community and expression. It’s a framing that exploits legitimate fears to protect trillion-dollar business models.

Lennon Torres knows both sides of this equation intimately. As a young dancer who gained national visibility on shows like Dance Moms, she built a following of hundreds of thousands on social media, and with that visibility came death threats about her queer identity and targeting by predatory adults through predatory product features. Now a key leader at the Heat Initiative and co-author of a piece in The Atlantic with Jonathan Haidt examining how tech companies use LGBTQ+ concerns as a regulatory shield, Lennon brings so much credibility to unpacking this manufactured divide and I am so grateful that she agreed to chat with me!

Subscribe to Lennon’s substack:

The Diary of a Palatable Trans Woman

Hey! How’d you find my diary? Well, since you’re here… welcome to Diary of a Palatable Trans Woman. This is where tea is spilt — sometimes with grace and occasionally all over the floor. Essays, culture critiques, and a lot of "did she just say that?"

By Lennon Torres

The entire video is shared above, and I’ve pulled some key insights below.

If you can’t watch the entire thing, here is a guide so you can jump to what you’re most interested in:

• 00:00 - Introduction to Lennon and the discussion

• 02:23 - How Kelly and Lennon met

• 03:07 - Lennon’s background as a child performer and early experiences online

• 05:01 - Holding dual realities: finding community and experiencing harm

• 07:45 - What unsafe online interactions look like (content warning: grooming, CSAM)

• 13:33 - Big Tech funding opposition to KOSA

• 15:00 - Senator Blackburn’s 2022 comments and original concerns about the bill

• 17:15 - The FTC workshop conflict and showing up for bipartisan work

• 22:01 - How KOSA evolved from content to design focus

• 24:30 - Fears about giving power to a hostile administration

• 27:31 - Kelly’s experience at Meta Horizon Worlds

• 28:45 - How Big Tech uses LGBTQ+ community as a shield

• 34:00 - Tim Cook vs. Mark Zuckerberg avoiding accountability

• 36:01 - Meta’s recent policy changes as a favor to Trump

• 37:00 - How to differentiate genuine concerns from industry propaganda

• 38:14 - Why Lennon broke up with her iPhone

• 46:11 - Comments from Aaron and Ava on platform dynamics

• 49:06 - Sylvia Rivera, palatability, and burnout

• 53:06 - How to reclaim the narrative

• 57:36 - Messages to lawmakers and Big Tech employees

Please, please, please subscribe to Lennon’s Substack,

Content vs. Design: How the Bill Has Evolved

The fears around KOSA aren’t unfounded. Early versions of the bill did include content-focused provisions that could have been weaponized to censor LGBTQ+ expression. And Senator Marsha Blackburn, one of the bill’s champions, made a 2022 comment in a YouTube video about protecting kids “from the transgender.”

The political landscape adds to the unease. The FTC led a workshop on social media harms to young people last summer, and they were running a workshop the following week on the dangers of gender-affirming care for minors.

As Lennon put it:

“What a slap in the face because I show up every day to advocate for all children. And I sit across the table from… Blackburn staff. And I go to her office. I shook her hand at an event last June. And I respect the work that she’s done to protect kids online. And it may be hard for some folks to hear that. And trust me, I feel you because she’s not traditionally an LGBTQ+ ally. And for that, I won’t sit down and have dinner with her. But I will show up to the meeting that is supposed to be about protecting all kids online.”

What Lennon points to is how the bill changed. The content provisions were removed. The Senate version that passed 91-3 focuses on design, targeting the predatory features and engagement-maximizing systems that harm kids, not the content they access. Even Blackburn compromised:

“She was willing to let go of some really, in my opinion, outdated and bigoted points of view to allow the bill to move forward because she realized that a design bill [like the current Senate version of KOSA] would help all kids. And therefore, it was in everyone’s best interest to pass it because we all want to protect kids.”

Some argue they supported the bill under the Biden administration but fear giving power to a government now hostile to queer people.

Lennon’s response:

“I think the administration that is in place right now is going to continue to harm LGBTQ+ people no matter what we do with Big Tech regulation... But at the same time, I totally hear the narrative that you don’t want to give the government any more power. But when I look at a design bill like KOSA, I don’t see it giving the government power. I see it holding the Big Tech companies accountable for the wrongdoing and the ability for a government, just or unjust, to take action against those companies to make their products safer for kids.”

“The bill can only do what the bill says.”

During confusing debates over the House version, Lennon called attorney Laura Marquez-Garrett in a moment of doubt:

“I need you to talk me down because I’m starting to believe it. I’m starting to believe the noise, like walk me down because I’m losing it… I’m afraid of the government too. I’m not invincible. Holy crap, what are we doing? Is this the right thing?”

What grounded her was going back to the text itself:

“They pulled up the bill and they started showing me and they said, this is what it says in the bill. ‘This is why they couldn’t do this’ There was no discrepancy. I wasn’t reading someone else’s analysis of it or I wasn’t asking a lobbyist what they think of the bill.”

Following the Money

Meta has spent over $50 million lobbying against KOSA. And they’ve found willing messengers in the very communities most harmed by their platforms.

“Big Tech knows that they can sell people who are struggling, who need community, who need connection. Oh, well, we’ve got it. We’ve got it and we’re going to give it to you with with sugar before and sugar after so it tastes really good and you’re going to want to come back for more.”

“Mark Zuckerberg’s passive income is our active scrolling. Like that is legitimately the business that they have created is the business of mining our attention.”

The exploitation runs deeper than advertising:

“They can mobilize an entire base that has the infrastructure to be loud and to break through noise because we’ve had to over and over and over again. And it makes me really... I was going to say sad, but today just really pisses me off that they will use that advantage to spark fear and spark some reactions. And, and then it just explodes because they’re prepped and ready...”

Lennon points to the absurdity of watching Tech CEOs fight each other while avoiding accountability:

“Tim Cook put a call in to the governor of Texas like ‘stop the App Store Accountability Act’ and was pushing forward KOSA the next day. And Meta endorsed the App Store Accountability Act [but lobbies against KOSA]... and it’s it’s like Tim Cook up here Mark Zuckerberg up here and they’re just yelling at each other, ‘you take responsibility!’ ‘no you take responsibility!’ ‘no!’ and it’s like, grow up and use the money you’re sitting on and pick yourself up and do the work.”

The Cost of Showing Up

Lennon is clear about what it means to be a trans woman advocating alongside legislators who have historically opposed LGBTQ+ rights:

“I’m not dying to go knock on Marsha Blackburn’s door. Like, I’m just not. And that’s OK. But I’ll still go to the meeting and I’ll still do the work because that’s what matters. And I have the capacity and the safety to do so.”

She doesn’t ask everyone to make the same choice:

“I want to remind you that your safety and your well-being should be the number one priority. And if that looks different than the choices that I’ve made, I respect it. And I want you to make those choices... we are in a very hard time and we’ve got to take care of ourselves and queer people need to be reminded sometimes that they have the agency to protect themselves a lot of the time. But sometimes they don’t. And to those people, I want to say stay strong and stay fighting because unless I disappear into the ether, I promise you there’s at least someone out there working for you.”

For those trying to sort through the noise, she offers a simple framework:

“Watch what people do. That’s how they feel. Watch what people do, not what they say. People can say a lot of crap.”

“If these people cared about kids, their actions would be so different. And that’s all I want people to pay attention to.”

Lennon’s Calls to Action

To lawmakers:

“Get to work, please get to freaking work, get to work, get to work, come on, like we work for the people, get to work, and stop taking money from Big Tech... It will come back and it will haunt you and stop the cycle. Just don’t do it. Don’t take money from Big Tech. Just don’t do it.”

To Big Tech employees:

“This isn’t about you. It’s about the system that you’re a part of. And if you know and are continuing to do harm, that’s on you, dog, and you got to leave.”

To allies:

“Be the one that raises your hand and says, ‘what is this going to do to the queer community?’ ‘Have we had a queer person in this room?’... be loud and raise your hand and like ask the questions because sometimes we get tired.”

Thank you , , , , , and many others for tuning into my live video with !

Shoutout to Aaron’s incredible podcast unpacking many of the various strategies of Big Tech’s shields at Superhuman Podcast

Superhuman Podcast

Companion Proof Hub to the Superhuman podcast—documents, timelines, transcripts, and expert notes on how platforms became legally “superhuman,” plus concrete steps to protect kids and fix the system.

By Aaron Ping