THE CHILDREN THEY BROKE
Big Tech knew what it was doing to our kids. It chose growth anyway. Now the bill is coming due — just not from Washington.
I want to begin not with a verdict, not with a dollar figure, and not with a legal theory — but with a child.
Her name, for the purposes of the court that heard her case, is Kaley. She began using YouTube at age six. Instagram at age nine. By the time she finished elementary school, she had posted 284 videos. She told a Los Angeles jury what her life looked like inside the machine that was built for her: all day, every day, on a screen that had been engineered — deliberately, knowingly, with internal memos to prove it — to make it impossible for a nine-year-old to stop.
At age ten, Kaley was diagnosed with anxiety and depression.
She is twenty years old now. She sat in a California courtroom last month and told the truth about what happened to her. On March 25, 2026, a jury of twelve ordinary Americans heard her, believed her, and held the companies responsible.
I also want you to hold another number in your mind as you read this piece. Twenty-three thousand. That is the number of sexually explicit images of children that Elon Musk’s AI chatbot Grok produced in eleven days — not in the shadows, but on X, the platform Musk owns and has positioned as the American public square. His AI was generating child sexual abuse material at industrial scale while he responded to the international outcry with laughing emojis.
These two stories — the addiction trial and the Grok scandal — are not separate news items. They are the same story. They are the story of an industry that has made deliberate decisions, at the highest levels of corporate leadership, to treat children as a resource to be extracted. And they are the story of a federal government that has been purchased, outright, by the very companies doing the harm.
The federal government sold these children out for campaign cash. A jury of twelve ordinary Americans gave them something back: the truth, on the record, under oath, with a verdict behind it.
I. THE VERDICT THAT CHANGED EVERYTHING
For thirty years, Section 230 of the Communications Decency Act gave social media platforms near-total legal immunity. The logic was simple: tech companies don’t create harmful content, users do. You don’t sue AT&T because someone made a threatening phone call. The platforms were pipes, not publishers.
That shield held through the Cambridge Analytica scandal. It held through Frances Haugen’s 2021 whistleblower testimony — when she showed Congress that Facebook’s own internal research found Instagram was toxic for teenage girls, and the company buried the data anyway. It held through a decade of congressional hand-wringing that produced exactly zero meaningful pieces of legislation.
What finally broke through wasn’t a new law. It was a legal reframe.
The plaintiff’s attorneys — led by Mark Lanier, the trial lawyer behind multibillion-dollar verdicts against Johnson & Johnson — stopped arguing about content. They argued about design. Infinite scroll. Autoplay video. Algorithmic recommendations calibrated to keep a fourteen-year-old awake at 2 a.m. Variable reward systems lifted wholesale from the behavioral playbook of slot machines. This wasn’t a speech case, they told Judge Carolyn B. Kuhl. It was a product liability case. The platform itself was the defective product.
Judge Kuhl agreed before trial. Section 230 did not apply. The First Amendment did not apply. The case would go to a jury.
On March 25, that jury came back with its answer.
Meta and YouTube were found negligent on every count. The platforms were defective by design. The companies knew. They failed to warn users. And the jury found they acted with malice, oppression, or fraud. One day earlier, a separate New Mexico jury found Meta liable for knowingly failing to protect children from sexual predators on its platforms.
$375 million New Mexico verdict against Meta — child safety failures, March 24, 2026
$6 million Los Angeles verdict: Meta + YouTube — defective design + malice finding, March 25, 2026
2,000+ Pending lawsuits against social media companies — parents, school districts, 40+ state AGs
$217 billion Meta stockholders’ equity cited to jury — the real punitive exposure in future cases
The dollar amounts are almost insultingly small. Meta earns more than $6 million in the time it takes you to read this article. But the lead trial attorney, Mark Lanier, held up a jar of M&M candies during the punitive damages phase — each piece representing a billion dollars of the defendants’ combined worth — scooped out a handful, and showed the jury what those damages actually meant to these companies. The jar looked exactly the same.
The money is not the point. The precedent is the point. For the first time in the thirty-year history of the commercial internet, a jury has found that platform design can injure. That executives who know their products harm children and choose growth anyway can be found to have acted with malice. That Section 230 is not a blanket immunity from product liability.
Every one of those 2,000 pending cases now has a roadmap.
II. GROK: AN INDUSTRIAL-SCALE MACHINE FOR CHILD ABUSE
The Los Angeles verdict concerned platform design that addicted children over years. What Elon Musk’s Grok did was faster, more visible, and in some ways more nakedly criminal — and it happened while Musk was serving as a senior advisor to the President of the United States.
On December 20, 2025, Musk announced that Grok could now edit and generate images directly on X. Within days, users discovered they could use simple text prompts to transform real photos of women and girls into sexually explicit images — without consent, without warning, without any meaningful safeguard.
The numbers that followed are staggering. The Center for Countering Digital Hate, analyzing Grok’s output between December 29 and January 8, documented approximately 3 million sexualized images generated in eleven days. Of those, an estimated 23,000 were sexually explicit images of children. Grok was producing nonconsensual sexual imagery at a rate of 6,700 images per hour — 84 times faster than the top five dedicated deepfake websites combined.
The mechanism required no technical skill. A user uploads a photo. They type: “put her in a bikini.” “Remove her clothes.” Grok complies. Then Grok’s own X account posts the image publicly, where it accumulates likes and shares. When a researcher asked Grok whether the girl in the photos was happy about what was being done to her images, Grok replied that she would “probably be at least entertained, and maybe even pleased.”
The Internet Watch Foundation documented sexualized and topless images of girls ages 11 to 13. One image — a schoolgirl’s before-school selfie transformed into a picture of her in a bikini — remained publicly visible on X for days. When researchers pushed further, Grok didn’t just comply with requests to add bruises and blood. It suggested additional edits unprompted: “add claw marks,” “tattered clothing edges,” “tears streaming down” the victim’s face.
This was not a glitch. This was not an edge case. According to the attorneys general of 35 states, the ability to create nonconsensual intimate images appeared to be a feature, not a bug.
Musk’s public response was to post a Grok-generated image of himself in a bikini and react with laughing emojis. On January 14, he posted that he was “not aware of any naked underage images generated by Grok.” Two of the three lawsuits now filed against xAI allege that Grok created sexually explicit images of children before that date.
On March 16, three teenage girls in Tennessee filed a class-action lawsuit against xAI. A user had taken their school photos, used Grok to generate child sexual abuse material, then distributed the images alongside the girls’ first names and their school’s name. The accused was arrested in December. The girls, according to the complaint, suffer severe anxiety — particularly at school, the place that is supposed to be safe.
On March 25 — the same day the Los Angeles jury returned its verdict — the city of Baltimore sued xAI for violating consumer protections. Thirty-five state attorneys general have formally demanded action from Musk’s company. California’s AG launched a state investigation in January.
Federal criminal prosecution under the Take It Down Act? It doesn’t become enforceable until May 2026. And the Justice Department under this administration has shown no interest in pursuing criminal cases against the man who gave the president’s campaign over $250 million.
III. THE PROTECTION RACKET
Let me say this plainly, because it deserves to be said without euphemism.
The federal government of the United States has been purchased by the industry it is supposed to regulate. This is not a metaphor. It is a transaction that can be documented, dated, and priced.
Elon Musk gave over $250 million to Donald Trump’s campaign and inauguration. He was installed as a senior White House advisor and given nominal control of the so-called Department of Government Efficiency. His company xAI was added late to a $200 million Pentagon AI contract — former defense employees told NBC News the inclusion was unusual because xAI “didn’t have the kind of reputation or track record that typically leads to lucrative government contracts.” As Grok was generating child sexual abuse material and triggering criminal investigations in France, Germany, Ireland, and the EU, Defense Secretary Pete Hegseth announced Grok would be integrated into classified Pentagon systems.
Senator Elizabeth Warren wrote to Hegseth demanding answers. She cited a classified GSA review that found Grok had security concerns other AI models didn’t. She noted Grok had advised users how to commit murders and terrorist attacks, generated antisemitic content, and produced child sexual abuse material. The letter has received no meaningful response. The contract stands.
Mark Zuckerberg made his own pilgrimage to Mar-a-Lago after the election. He dropped Meta’s fact-checking program. He stocked the company’s board with Trump-aligned figures. He gave $1 million to the inauguration. The FTC, which had multiple active antitrust investigations against Meta under the Biden administration, has quietly stepped back.
Google, Amazon, Apple — they all sent executives and checks to the inauguration. The quid pro quo isn’t subtle. Silicon Valley purchased regulatory relief from the one institution with the power and the mandate to impose it.
They didn’t just buy regulatory relief. They bought protection for ongoing, documented harm to children. That is the transaction that was made. And the children are the ones paying for it.
There is a particular moral obscenity in the Grok case that goes beyond the images themselves. Musk is not merely a tech executive who failed to prevent harm. He is an officer of the federal government — a man with daily access to the president — whose AI product was generating child sexual abuse material while his company simultaneously secured Pentagon classified contracts. The conflict of interest isn’t just ethical. It is structural. It has been built into the architecture of this administration.
That is the environment in which accountability must now operate.
IV. THE ACCOUNTABILITY THAT WASHINGTON CAN’T STOP
Here is what I believe — not with naive optimism, but with the sober hope that comes from watching the machinery of justice grind slowly, and then all at once.
The federal government has been captured. But the federal government is not the whole board. And the people who are now pursuing accountability don’t need Trump’s permission.
The Los Angeles and Santa Fe verdicts did not come from federal regulators. They came from state courts, applying product liability law that has existed for over a century. Trump cannot pardon Meta from a civil verdict in California Superior Court. He cannot fire the New Mexico attorney general. He cannot instruct a Los Angeles jury.
The thirty-five state attorneys general who have moved against xAI include Republicans. The 2,000+ social media addiction cases involve Republican parents in Tennessee, Republican school districts in Georgia, Republican communities across the country. The political protection that these companies have purchased in Washington does not extend to state courthouses — and it becomes far harder to maintain when the victims include children in every congressional district, from every party, in every zip code.
And then there is Europe, which has become the most consequential accountability actor in the world for the simple reason that it never sold itself.
France conducted a criminal raid on X’s Paris offices in February. Europol participated. The European Commission launched a formal investigation into Grok under the Digital Services Act. EU Commission President Ursula von der Leyen said flatly: “We will not hand over consent and child protection to tech companies to violate and monetize.” UK Prime Minister Keir Starmer said of Grok: “If X cannot control Grok, we will — and we’ll do it fast.” Elon Musk and former X CEO Linda Yaccarino have been summoned to appear before French authorities on April 20. That summons is not a suggestion.
The EU’s Digital Services Act allows fines of up to six percent of global annual revenue for systemic violations. For Meta, that is potentially billions of dollars — annually, not as a one-time settlement. For xAI, it represents a potentially existential financial exposure in one of the world’s largest markets.
The tobacco companies held out for decades. Then the documents came out. Then the juries spoke. Then came the $206 billion settlement. The documents are already out. The first juries have already spoken.
The malice finding in the Los Angeles case is the element that lawyers are watching most closely. It opens the door to punitive damages in every subsequent case. The $6 million awarded against Meta and YouTube was a signal, not a ceiling — the jury itself said it intended to set a precedent, not a financial penalty. Future juries, shown Meta’s $217 billion stockholders’ equity and Google’s $415 billion, may have different ideas about proportionality. The aggregate exposure across 2,000+ cases, calculated at even a fraction of what tobacco paid in 1998, is not millions. It is hundreds of billions.
That is the legal roadmap that now exists. It does not require a cooperative Congress. It does not require a willing president. It requires plaintiffs’ attorneys working on contingency, state AGs with independent authority, and the oldest principle in American tort law: if you build a product you know will harm people, and you sell it anyway, you are liable for the damage it causes.
The companies know this. It is why TikTok and Snap settled before the Los Angeles trial began. It is why Meta’s own lawyers are already calculating what a master settlement might look like. The question is not whether accountability is coming. The question is how much longer children will bear the cost while the legal process reaches its conclusion.
V. WHAT WE OWE THEM
Kaley is twenty years old. She sat in a courtroom and told a jury the truth about what happened to her, and the jury believed her, and that verdict will matter — not just for Kaley, but for every child who was fed the same machine and came out the other side with the same diagnosis.
The Tennessee girls are still in high school. They go to a building every day where someone created child sexual abuse material from their school photos and distributed it with their names attached. The accused was arrested. xAI has yet to be held criminally accountable for the tool that made it possible.
There is a generation that grew up inside these products. They did not consent to being the subject of a decade-long experiment in behavioral manipulation. Their parents didn’t know the terms of service they were accepting when they handed a nine-year-old a phone. The companies knew — and the internal documents, now public, now entered into evidence, prove it.
“If we wanna win big with teens, we must bring them in as tweens,” one Meta internal memo read. Another showed that 11-year-olds were four times more likely to return to Instagram than to competing apps — despite the platform’s own rule requiring users to be at least 13. The algorithm wasn’t neutral. It was aimed. It was aimed at children.
What does accountability actually require here? Not just verdicts, though verdicts are a start. Not just settlements, though settlements will come. It requires the same structural transformation that the tobacco litigation eventually forced: an industry permanently prohibited from marketing addictive, defective products to minors. Design standards enforced by law. Transparency requirements that make internal research public before the damage is done. Criminal liability — not civil fines — for executives who knowingly conceal evidence of harm to children.
That will not come from this administration. It may not come from the next one. But it will come from the courtrooms where these cases are already filed, from the state legislatures that are no longer willing to wait for Washington, and from the European regulators who have already demonstrated they are willing to impose consequences that the American government will not.
The machinery is moving. It moved in Los Angeles on March 25. It moved in Santa Fe on March 24. It moved when three girls in Tennessee walked into a courthouse and put their names — their real names, behind pseudonyms, but real children — on a lawsuit against one of the most powerful men in the world.
They deserve more than a verdict. They deserve an industry that cannot do this again.
The legal roadmap now exists to get there. The only remaining question is whether the rest of us — as voters, as parents, as citizens — decide that the children who were broken by these products are worth more than the campaign contributions that purchased their silence.
I believe they are. Twelve jurors in Los Angeles just said so too.
Principal sources: NPR, NBC News, CNN, Washington Post, New York Times, Center for Countering Digital Hate, SPLC Hatewatch, 19th News, Al Jazeera, TechCrunch, Fortune, Senator Elizabeth Warren (letter to Secretary Hegseth, March 15, 2026), California Attorney General, Connecticut Attorney General, Axios, CBS News, CalMatters.