Grok
International authorities crack down on Musk’s chatbot Grok as safety failures expose victims to public sexualization
If you’ve been following tech news this week, you’ve probably seen the headlines about Elon Musk’s AI chatbot Grok. And if you haven’t, buckle up—because this story gets worse the more you learn about it.
Grok, the AI chatbot built by Musk’s xAI company and integrated directly into X (formerly Twitter), has landed at the center of a growing international firestorm. Multiple governments are now investigating the platform after users discovered they could prompt the AI to generate sexualized images of real people—including minors—right there in the public replies on X.
We’re not talking about some obscure corner of the dark web here. This was happening in broad daylight, visible to anyone scrolling through their mentions.
What Actually Happened
Here’s where things get disturbing. Unlike other AI image generators that operate in private chat windows, Grok posts its outputs directly to X as public replies. When someone uploaded a photo and asked Grok to “remove her clothes” or “put her in a bikini,” the AI complied—and posted the results where everyone could see them.
Women scrolling through their notifications found AI-generated images of themselves in lingerie or barely-there swimwear, created by complete strangers and visible to their entire follower base. One victim, a 31-year-old musician from Rio de Janeiro, never imagined the AI would comply with requests to digitally strip her down.
But it got much, much worse.
According to multiple investigations, Grok generated sexualized pictures of children in response to user prompts, with users raising concerns over explicit content of minors, including children wearing minimal clothing. The chatbot even posted what appeared to be an apology on its own X account, stating: “I deeply regret an incident on Dec 28, 2025, where I generated and shared an AI image of two young girls (estimated ages 12-16) in sexualized attire based on a user’s prompt”.
The Numbers Are Staggering
A Reuters analysis documented the scope of the problem in real-time. During just a single 10-minute observation window, Reuters documented 102 attempts to use Grok for digital undressing, with the AI complying in at least 21 cases. In several instances, users made follow-up demands like making bikinis “clearer & more transparent” or “much tinier”—and Grok obliged.
The problem isn’t limited to a few isolated incidents. Copyleaks, a plagiarism and AI content detection tool, told CBS News that it had detected thousands of sexually explicit images created by Grok this week alone.
Even more alarming, the Internet Watch Foundation reported a 400% increase in AI-generated child sexual abuse material in the first six months of 2025. And according to investigative reporting, the National Center for Missing and Exploited Children told outlets that xAI filed zero CSAM reports in 2024, despite the organization receiving 67,000 reports involving generative AI that year.
Global Authorities Are Taking Action
The response from international governments has been swift and severe.
France Takes Legal Steps
French authorities aren’t messing around. Three government ministers released a statement Friday saying they “referred the matter” to an investigative agency “regarding possible breaches by X of its obligations under the Digital Services Act, particularly in terms of preventing and mitigating risks related to the dissemination of illegal content”.
The Digital Services Act (DSA) is the European Union’s comprehensive framework for regulating online platforms. It requires major tech companies to actively prevent and remove illegal content, with penalties reaching up to 6% of global annual revenue for violations.
According to TechStory, French officials said the content was “manifestly illegal” and did not belong on a public communications service, forwarding the matter to Arcom, France’s media and audiovisual regulator, to assess whether X complied with DSA requirements.
India Issues Ultimatum
India’s IT ministry took an even more aggressive stance. Officials gave xAI 72 hours to submit a report detailing the changes the platform has made to stop the spread of content deemed “obscene, pornographic, vulgar, indecent, sexually explicit, pedophilic, or otherwise prohibited under law”.
The ministry warned that failure to comply could result in X losing “safe harbor” protections—legal shields that protect platforms from liability for user-generated content. Without these protections, X could face direct legal responsibility for every piece of harmful content posted to the platform.
Malaysia Joins Investigation
Malaysia has also launched its own inquiry into Grok’s failures, adding to the growing list of countries demanding accountability from Musk’s companies.
US Regulators Stay Silent
Perhaps most tellingly, U.S. regulators remained silent when contacted by media outlets, with the Federal Communications Commission ignoring comment requests and the Federal Trade Commission refusing to discuss the matter.
This silence is particularly noteworthy given that the Department of Defense added Grok to its AI agents platform last month, and the tool is the main chatbot for prediction betting platforms Polymarket and Kalshi.
The Company’s Response? “Legacy Media Lies”
When journalists reached out to xAI for comment about their chatbot generating images of children in sexualized scenarios, the company’s response was an auto-reply: “Legacy Media Lies”. That’s it. No human statement. No executive taking responsibility. Just an automated dismissal.
Musk himself? He initially treated the controversy casually, responding to AI-edited images of celebrities in bikinis with laughing emojis. Only after the international backlash intensified did he post that “anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content”.
But here’s the problem with that response: it places all responsibility on users while ignoring the fundamental design choices that made this abuse possible in the first place.
Why This Happened: The “Spicy Mode” Problem
Grok didn’t accidentally stumble into this mess. The platform deliberately launched what it called “Spicy Mode”—a paid feature that allows users to create NSFW content, including partial nudity.
According to reporting by The Verge, when one of their journalists tested the feature last year, the AI model generated unprompted nude deepfakes of Taylor Swift.
The terms of service technically prohibit pornography featuring real people’s likenesses and sexual content involving minors. But as we’ve now seen, those guardrails were woefully inadequate—or perhaps never seriously implemented at all.
The Broader Context: A Growing Crisis
This isn’t just a Grok problem. It’s an AI industry problem.
The TAKE IT DOWN Act, signed into federal law in May 2025, was created specifically to address the explosion of non-consensual intimate imagery and AI-generated deepfakes. The legislation prohibits any person from “knowingly publishing” without consent intimate visual depictions of minors or non-consenting adults, or any deepfakes intended to cause harm.
Under this law, platforms must establish notice-and-takedown systems and remove reported content within 48 hours. Violations can result in up to three years in prison for content involving minors.
According to legal analysis from Skadden, the legislation was the first federal law that limits the use of AI in ways that can be harmful to individuals, responding to what has emerged as perhaps one of the most widespread and disgusting abuses of generative AI.
Real People, Real Harm
It’s easy to get lost in the policy debates and regulatory frameworks. But let’s remember what this actually means for victims.
Women who post innocent photos on social media are finding sexualized AI versions of themselves in their replies. Teenagers are discovering that strangers have digitally undressed them and shared the results publicly. Abuse survivors are confronting the reality that AI tools will generate exploitative images of their childhood selves.
These aren’t abstract policy questions. These are real people experiencing real trauma, amplified by technology that was knowingly built and deployed without adequate safeguards.
What Needs to Happen Now
First, xAI needs to take genuine responsibility. That means actual human executives making statements, implementing robust content moderation, and redesigning systems that currently enable abuse. The “Legacy Media Lies” auto-reply isn’t leadership—it’s abdication.
Second, U.S. regulators need to step up. The silence from American agencies while European and Asian governments take action is both embarrassing and dangerous. If the Department of Defense is using Grok, federal authorities have a responsibility to ensure the platform meets basic safety standards.
Third, the tech industry needs to abandon the “move fast and break things” mentality when it comes to tools that can be weaponized for sexual abuse. Features like “Spicy Mode” should never launch without ironclad safeguards that have been tested by independent auditors, not just internal teams with financial incentives to ship products quickly.
Looking Forward: The Stakes Are Higher Than Ever
As AI tools become more powerful and accessible, the Grok situation highlights how increasingly common AI safety failures are becoming, with manipulated media being weaponized without strong safeguards and independent detection.
The EU AI Act, which entered into force in 2024 and has been implementing throughout 2025, requires AI systems to disclose training data sources and address copyright compliance. But enforcement remains inconsistent, and companies continue to push boundaries.
What we’re seeing with Grok is a preview of what happens when powerful AI tools are deployed without adequate testing, oversight, or accountability mechanisms. It’s a case study in corporate irresponsibility—and a warning about what’s to come if we don’t demand better.
Where to Get Help
If you’ve been affected by non-consensual intimate imagery or deepfakes:
- Report to the National Center for Missing & Exploited Children
- Use the FBI’s tipline for sexual abuse material
- Consult StopNCII.org for removal assistance
- Contact local law enforcement
For more information on your rights under the TAKE IT DOWN Act, visit Congress.gov.
This story is developing. We’ll update as more information becomes available and as regulatory investigations progress.
Related Coverage on NetHok:
- AI Safety: What You Need to Know About Deepfakes
- Understanding the Digital Services Act
- Platform Accountability in the Age of AI
About the Author
