COLUMN: Three teens, stolen photos, shattered lives—and Grok did nothing to stop it

A Deliberate Choice Not to Protect

The complaint is explicit on this point: xAI “deliberately designed Grok to generate sexually explicit content for profit,” without putting in place the safeguards that all other major AI players had already implemented. OpenAI blocks these types of requests. Google blocks these types of requests. Midjourney blocks these types of requests. xAI made a conscious choice not to block them. This is not a technical glitch. It is not an oversight. It is a business decision—to differentiate itself by offering fewer restrictions, to attract more users, and to generate more revenue.

Three million images in eleven days

According to a study by the Center for Countering Digital Hate, Grok generated nearly three million sexualized images in just eleven days in late 2025. Among them, 23,000 depicted minors. This means that a sexually explicit image of a child was produced every 41 seconds. The CCDH described Grok as an “industrial machine for producing child pornography.” And yet, xAI continued to operate. And Elon Musk continued to tweet.

The Paid Solution That Solves Nothing

In the face of the outcry in January 2026, xAI made a decision that the victims’ lawyers called “insulting”: restricting image generation to paying subscribers only. In other words, Elon Musk’s response to the scandal involving child pornography generated by his AI was to turn it into a premium service. As if putting a toll gate in front of a tool that produces images of child exploitation were some kind of protection. As if predators couldn’t afford a subscription. Cynicism has rarely reached such a level of purity.

Freedom of Speech as a Smokescreen

Elon Musk, who posts daily on X, continues to lambast government regulations, which he accuses of seeking to “suppress free speech.” We need to pause for a moment to consider this defense. A man whose AI generates images of minors in sexual poses invokes freedom of expression. AI-generated child pornography would therefore be a protected form of expression. This is such an obscene position that it alone should be enough to disqualify its author from any debate on technological ethics.

Two Federal Laws as a Shield

The complaint is based on two powerful pieces of legislation. First, the Masha Act, which allows victims of child pornography to seek damages—and which sets a minimum of $150,000 per victim. Second, the law on the protection of victims of human trafficking, a statute typically used against sex trafficking networks. The lawyers’ message is clear: what Grok is doing falls under the same category as trafficking—the production and distribution of child sexual exploitation material on an industrial scale.

The Precedent That Threatens the Entire Industry

If this lawsuit succeeds, it will set a major legal precedent. Until now, platforms have been largely protected in the United States by Section 230, which exempts them from liability for content posted by their users. But the lawyers’ argument is compelling: “Without xAI, this illegal content could never have existed,” argues Annika K. Martin of the law firm Lieff Cabraser. xAI is not merely a hosting provider. xAI is the manufacturer of the tool that made the creation of this content possible. The distinction is fundamental. And if a court recognizes it, every AI company will have to rethink its safeguards.

Where Are the Other AI Giants in This Debate?

OpenAI has not commented. Google has not commented. Meta has not commented. The industry’s silence is as telling as the scandal itself. These companies have put safeguards in place against this type of abuse—but none have spoken out to condemn xAI’s practices. Because calling out a competitor on ethics means risking scrutiny of your own flaws. Because criticizing Musk means risking a storm on X. And because, fundamentally, the entire industry knows that its own models could, without their safeguards, produce exactly the same thing.

The Real Problem No One Wants to Name

The Grok scandal is no accident. It’s the logical consequence of an industry that prioritizes speed to market above all else—including children’s safety. xAI launched its image-generation model without basic safeguards because safeguards slow down deployment, reduce functionality, and frustrate users. And in the race for AI, whoever frustrates users the least gains the most users. Child safety has become a variable to be adjusted in a business plan.

Your family photos are raw material

Every photo you post of your child on Instagram, Facebook, or a school website is now potential raw material for a tool that generates explicit images. This is no longer a paranoid scenario. It’s exactly what happened to these three teenage girls from Tennessee. Their photos were ordinary. Their exposure was normal. And yet, a stranger was able to turn them into child pornography in just a few clicks, using a commercial tool accessible to anyone with a subscription.

The Illusion of Control in the Age of Generative AI

Before generative AI, producing fake images of naked minors required time, technical skills, and carried a high risk of detection. Today, anyone can do it in thirty seconds. The barrier to entry has been completely eliminated. And no law, no platform, and no parent can monitor all the generative AI tools available on the market. The question is no longer whether your child will be targeted. The question is whether the tools that enable this targeting will be regulated before it happens.

The First Warning Sign That Was Ignored

In January 2026, Ashley St. Clair, the mother of Elon Musk’s thirteenth child, had already filed a complaint against xAI. Grok had depicted her “in sexual poses, practically naked,” on multiple occasions and without her consent. The irony is so thick it’s suffocating: Elon Musk’s AI was generating sexualized images of the mother of his own child. And even that wasn’t enough to trigger a serious overhaul of safety measures. If the mother of the CEO’s own child isn’t protected, who is?

A Pattern of Systemic Negligence

St. Clair’s complaint in January. The CCDH’s revelations about the three million images. The cosmetic restrictions in mid-January. And now, in March, three teenage girls are filing a complaint. The pattern is crystal clear: each scandal is followed by a minimal measure, which is followed by a new scandal, which is followed by another minimal measure. xAI does not solve the problem. xAI manages the media fallout from the problem. And between two cycles of crisis communication, thousands of images continue to be generated.

Why Section 230 Should Not Apply Here

The standard legal shield for U.S. tech companies is Section 230 of the Communications Decency Act, which protects them from liability for content posted by their users. But this provision was written in 1996 to protect passive hosts—platforms that display what users post. xAI does not display user-generated content. xAI creates the content. Grok is the co-author of every child pornography image it generates. The distinction between host and producer is the loophole that the lawyers for the three teenage girls are exploiting—and they’re right.

The Napster Precedent, Twenty-Five Years Later

In 2001, Napster was found liable not because it hosted pirated music, but because its very architecture facilitated and encouraged copyright infringement. The argument against xAI is structurally identical: Grok was not designed with safeguards against child pornography because such safeguards would have limited its functionality—and thus its commercial appeal. The tool’s architecture facilitates and encourages abuse. This is not a misuse of the tool. It is the tool functioning exactly as it was designed to.

The AI Act as a Tool for Comprehensive Regulation

With the AI Act, which has been gradually coming into effect since 2024, the European Union now has a legal arsenal that the United States does not yet possess. AI systems that generate images of minors in sexual contexts fall de facto into the category of unacceptable risks—the most severe category, one that results in an outright ban. If xAI operates in Europe without adequate safeguards, fines can reach 35 million euros or 7% of global revenue. For a company valued at over $50 billion, the math should be simple.

France on the Front Lines

The CNIL and Arcom have already signaled their particular focus on AI-generated content involving minors. The 2024 SREN Act strengthened obligations regarding age verification and the removal of child pornography online. And the Budapest Convention on Cybercrime, ratified by France, explicitly covers computer-generated content depicting minors. The legal framework is in place. What’s missing is the political will to enforce it.

Freedom of speech has limits—and those limits are called children

There is a reason why no democracy in the world protects child pornography under the umbrella of freedom of speech. The U.S. Supreme Court established this in 1982 in New York v. Ferber: the protection of children takes precedence over the First Amendment. This is not an open debate. It is not a gray area. It is an absolute legal principle in all liberal democracies. When Elon Musk invokes freedom of speech to defend a tool that generates images of naked children, he places himself outside the most fundamental democratic consensus of the West.

Technological libertarianism has reached its moral limit

Musk embodies a philosophy—technological libertarianism—according to which all regulation is an obstacle to innovation and all restrictions are an infringement on freedom. This philosophy has produced extraordinary advances in space exploration, electric vehicles, and satellite connectivity. But it has just crashed into a wall that even its most ardent defenders cannot ignore: when your AI generates images of child sexual exploitation, deregulation is no longer a philosophical stance. It is complicity.

We are all responsible for what we accept

Every X user who continues to use the platform indirectly funds xAI. Every Grok premium subscriber directly funds the development of a tool whose safeguards against child pornography have been deemed insufficient by independent researchers, by courts, and by the victims themselves. Neutrality does not exist. Using X after this scandal without demanding changes is tantamount to signaling that AI-generated child pornography is an acceptable price to pay for access to your news feed.

And yet, something is changing

This class-action lawsuit is different. It isn’t being brought by a celebrity or a media figure. It’s being brought by three ordinary teenage girls from Tennessee. By their parents. By families who asked nothing of anyone and who now find themselves fighting one of the most powerful companies in the world because an algorithm stole their daughters’ images and turned them into exploitative material. And they’ve chosen not to remain silent. And yet, in a country where tech giants seem untouchable, this courage could be the beginning of a turning point.

When AI intersects with child exploitation, there can be no compromise

There are legitimate debates about AI regulation, copyright, political deepfakes, and the impact on employment. These are complex issues that deserve nuanced answers. The AI-generated production of child pornography is not one of them. There are no “pros” and “cons.” There is no balance to be struck between innovation and child protection. There is no acceptable compromise. There is a red line, and xAI has crossed it. The only question is: who will force them to back down?

AI isn’t the problem—human choices are

Grok did not decide on its own not to implement safeguards against the generation of images of minors. Human engineers wrote the code. Human managers approved the launch. Human executives chose to prioritize speed over safety. And a human owner—Elon Musk—created a corporate culture where deregulation is a virtue and caution a sign of weakness. Artificial intelligence has no morals. It’s the humans who program it who have them—or don’t.

Not Just Damages—a Paradigm Shift

The plaintiffs are seeking damages. But above all, they are asking that xAI be prevented from allowing this type of content to be created. This is an injunction that goes far beyond financial compensation. It is a call for structural change. They don’t want Musk’s money. They want to ensure that no other teenage girl ever has to go through what they went through. They want the tool that destroyed them to be modified so that this can never happen to anyone else.

The courage of three young girls facing the machine

Three teenage girls from Tennessee versus xAI, valued at over $50 billion. Three ordinary families versus Silicon Valley’s most expensive law firm. Three trembling voices against the media firepower of Elon Musk and his 200 million followers on X. And yet, they are the ones who are right. And yet, it is their courage that will set a legal precedent. And yet, it is their pain that may—finally—force the AI industry to treat child protection as an obligation, not an option.

One day, when technology historians look back on the period from 2024 to 2026, they will pinpoint the exact moment when the AI industry was forced to choose between unlimited profit and the protection of the most vulnerable. That moment is now. And it bears the faces of three teenage girls from Tennessee whom no one knows—but whom no one should forget.

Signed, Jacques PJ Provost