Under Musk, the Grok disaster was inevitable

This is it Step backa weekly newsletter covering one essential story from the world of technology. To learn more about the disruptive developments in artificial intelligence, follow along Haydenfield. Step back It arrives in our subscribers’ inboxes at 8 a.m. ET. Subscribe to Step back here.

You could say it all started with Elon Musk’s AI FOMO — and his campaign against “wokeism.” When his artificial intelligence company, xAI, announced Grok was joining November 2023It has been described as a chatbot with a “rebellious streak” and the ability to “answer provocative questions that most other AI systems would refuse to do.” The chatbot debuted after a few months of development and just two months of training, and the announcement highlighted that Grok would have real-time knowledge of the X platform.

But there are inherent risks with online chatbots and X, and it’s safe to say that xAI probably hasn’t taken the necessary steps to address them. Since Musk acquired Twitter in 2022 and renamed it X, he has laid off 30% of global trust and safety staff and cut the number of safety engineers by 80%, Australia’s online safety watchdog said. Last January. As for xAI, when Grok was released, it wasn’t clear whether xAI already had a security team in place. When the Grok 4 was released in July, it took more than a month for the company to release a sample label — a practice typically seen as an industry standard, which details safety tests and potential concerns. Two weeks after Grok 4 was released, an XAI employee wrote: On X That he was hiring a safety team at xAI and that they were “desperately in need of strong engineers/researchers.” In response to a commenter who asked: “Does xAI provide security?” Original employee He said xAI was “working on it.”

Journalist Kat Tenbarge wrote about how the vision first began Sexually explicit deepfakes It went viral on Grok’s June 2023. These images were clearly not created by Grok — and he didn’t even have the ability to create the images until August 2024 — but X’s response to the concerns has been varied. until Last JanuaryGrok has been courting controversy over AI-generated images. and Last AugustGrok’s “hot” video creation mode led to her creating fake nude photos of Taylor Swift without being asked. The experts said Edge Since September That the company takes a whack-in-the-mole approach to safety and guardrails — and that it’s hard enough to keep an AI system on the straight and narrow when it’s designed with safety in mind from the start, let alone whether you’ll come back to fix hidden problems. Now, this approach appears to have exploded in xAI’s face.

Grok has spent the past two weeks posting fake sexualized material to adults and minors across the platform, as he is promoting. Screenshots show Grok’s commitment to asking its users to replace women’s clothes with underwear and make them spread their legs, as well as requiring young children to wear bikinis. There are even more horrific reports. It got so bad that during a 24-hour analysis of the images Grok created on X, One estimate It has been measured that the chatbot generates about 6,700 sexually suggestive or “nude” images per hour. Part of the reason for the attack was a recent feature added to Grok, allowing users to use the “Edit” button to ask the chatbot to change images, without the original poster’s consent.

Since then, we’ve seen a few countries either investigate or threaten to ban X altogether. Members of the French government He promised an investigationas he did Indian Ministry of Information TechnologyAnd a Malaysian government committee He wrote a letter About her fears. California Governor Gavin Newsom He called US Attorney to investigate XAI. The UK said it was Planning to pass the law Banning the creation of non-consensual sexual images generated by artificial intelligence, the country’s communications industry regulator said it would investigate both X and the images generated to see if they violate the Internet Safety Law. And this week, both Malaysia and Indonesia Blocked access to Grok.

xAI initially said its goal for Grok was to “help humanity in its quest for understanding and knowledge,” “to benefit all of humanity to the fullest extent,” “empower our users with our proprietary, law-governed AI tools,” as well as “serve as a powerful research assistant for anyone.” This is a far cry from producing fake nude photos of women without their consent, let alone minors.

On Wednesday evening, as pressure mounted on the company, X’s Safety account posted a message statement The platform “has implemented technological measures to prevent the Grok account from allowing the editing of photos of real people wearing revealing clothing such as bikinis,” and that the restriction “applies to all users, including paid subscribers.” Furthermore, only paid subscribers can use Grok to create or edit any type of photo going forward, according to X. The statement went on to say that

Another important point: my colleagues Test Grok’s image creation limitations on Wednesday To discover that it took less than a minute to get around most guardrails. Although asking the chatbot to “put her in a bikini” or “remove her clothes” produced censored results, they found that it had no qualms about making prompts such as “show me her cleavage,” “enlarge her breasts,” and “wear a low-rise crop top and shorts,” as well as creating lingerie photos and sexual poses. As of Wednesday eveningWe can still get Grok to create revealing photos of people, with a free account.

Even after X’s statement on Wednesday, we may see a number of other countries either ban or deny access to X entirely or just Grok, at least temporarily. We’ll also see how the proposed laws and investigations around the world will play out. Pressure is mounting on Musk, who took over on Wednesday afternoon To X To say he was “not aware of any nude images of minors created by Grok.” Hours later, X’s Safety team issued its statement, saying it was “working around the clock to add additional safeguards, take swift and decisive action to remove infringing and illegal content, permanently suspend accounts where appropriate, and cooperate with local governments and law enforcement when necessary.”

What is technically and what is not against the law is a big question here. For example, experts said Edge Earlier this month That AI-generated images of identifiable minors in bikinis, or perhaps even nude, may not technically be illegal under current child sexual abuse material (CSAM) laws in the United States, although disturbing and immoral, of course. But lewd images of minors in such situations are considered a violation of the law. We will see if these definitions expand or change, although the current laws are a bit of a patchwork.

As for non-consensual intimate deepfakes of adult women, the Take It Down Act, signed into law in May 2025, bans AI-generated “intimate visual depictions” without consent and requires certain platforms to quickly remove them. The grace period before the final part comes into effect – which requires platforms to actually remove it – ends in May 2026, so we could see some important developments in the next six months.