A new feature on Elon Musk’s Grok chatbot is facing intense backlash from governments, advocacy groups, and users worldwide. Grok, developed as Musk’s AI version of ChatGPT, and integrated into X, was marketed as a more “unfiltered” alternative to mainstream AI tools. While the chatbot launched in 2023, a major new feature was introduced at the end of 2025 — Grok Imagine, which was meant to allow users to create visuals from text prompts and animate images. X AI promoted the tool as a bold experiment in creative expression, a next-generation visual playground.

The idea, in theory, was straightforward — let users generate pictures and videos from prompts, remix existing images, and unleash creative potential in ways that normal text-based AI can’t.

However, it’s the option of enabling “Spicy mode,” which allows for nudity and creating adult content, that turned a user’s creative expression into a nightmare for unsuspecting victims. Grok Imagine’s X-rated levers became tools for abuse.

Pretty much immediately after the feature launched, users discovered a loophole where AI would comply with unfiltered prompts to remove clothing from anyone in a photo, and to carry out other inappropriate requests on any image it could find. The Associated Press described how a simple prompt could get Grok to undress someone’s social photo, generate sexualized imagery of their face and body without their permission, and post the output publicly — even if the photo had since been deleted from the original account. It didn’t take long for women to come forward with accounts of discovering they had been turned into the subjects of these images — often learning of the exploitation only after the pictures had already circulated to millions of social media users.

In chilling detail, The Free Press relayed how a real user, a young woman in the UK, saw Grok-generated content of her in digital “undress” on her own social feed, and how that “felt … as violating as if someone had actually posted a nude or a bikini picture of me.”

The Christian Post highlighted how safety advocates and women’s rights groups reacted with alarm to the bot’s ability to turn ordinary photos into sexually explicit deepfakes. Representatives of groups like the National Center on Sexual Exploitation (NCOSE) emphasized that Grok’s behavior wasn’t just a technical glitch; it was enabling horrifying violations of women’s privacy and safety.

Critics also pointed out that vague rules — such as Grok’s instructions to “assume good intent” when asked for images of “girls” or “teenagers” — left the system open to generating disturbing content that verges on federally illegal child sexual abuse material (CSAM).

In response to the problem, Musk stated that he was “working with local governments and law enforcement as necessary.” He warned, “Anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content.”

Governments and regulators around the world, however, are not waiting to take serious steps to protect their citizens.

In the UK, regulators launched a formal probe into whether Grok violated online safety laws. Tech Secretary Liz Kendall made it clear that tools used to create “non-consensual intimate images” should be treated as criminal, with serious fines or even platform bans on the table. At the EU level, the European Commission began examining whether Grok complies with European online safety rules and ordered X to preserve all internal data tied to the feature through 2026. France has since opened its own criminal investigation under the Digital Services Act, while officials in Poland and Brazil publicly called out Grok’s lack of safeguards, warning that digital exploitation doesn’t stop at national borders.

Elsewhere, the response was even tougher. India ordered X to take down unlawful Grok-generated content and submit to a rapid technical and governance review, giving the company just 72 hours to comply. Malaysia and Indonesia went a step further, becoming the first countries to block access to the AI tool altogether. Officials cited repeated misuse to create sexually explicit, non-consensual images of women and minors, saying the feature posed a direct threat to digital safety and that Grok won’t be welcomed back without major changes.

Across these jurisdictions, critics are united on one point: the ability to sexualize images of real people — especially young girls and women — without consent is not free speech but a profound violation of dignity and privacy.

Elon Musk seems to be working to fix the issue, using his company to restrict Grok Imagine’s image-generation and editing features to paying subscribers in a bid to curb rampant misuse, though this could be seen as monetizing abuse rather than a real solution.

However, X AI’s automated responses to media requests — some as simple as “Legacy Media Lies” — have also not helped public perception.

Put simply, Grok Imagine’s controversial rollout has sparked one of the most intense waves of global regulatory scrutiny of an AI feature to date. Governments are moving beyond investigations to real policy actions — criminalizing non-consensual deepfake imagery, issuing takedown orders, and, in some cases, temporarily banning access.

For many advocates and policymakers, this is a critical moment in AI ethics. It underscores just how quickly generative AI went from fun novelty to powerful tool for harm when safeguards are weak and incentives are misaligned.

As lawmakers, safety advocates, and tech leaders continue to grapple with these challenges, Grok Imagine’s saga will almost certainly be cited as a case study in how not to deploy sensitive AI features without ethical guardrails.