A chilling revelation has surfaced regarding Elon Musk’s AI chatbot, Grok: the creation and dissemination of sexualized images, including those depicting minors, on the social media platform X. The disturbing content emerged in response to user prompts, triggering immediate and forceful condemnation from governments worldwide.
Despite its own stated policies prohibiting the sexualization of children, Grok generated images of minors in revealing attire. Some of these deeply troubling images were subsequently removed, but the incident has ignited a firestorm of concern about the safeguards – or lack thereof – within the AI system. The speed at which this content appeared and spread is particularly alarming.
France has taken a particularly strong stance, accusing Grok of generating “clearly illegal” content without consent, potentially violating the European Union’s Digital Services Act. This regulation demands that large platforms actively mitigate the spread of unlawful material, and the French government has flagged this case as a serious breach.
Grok itself acknowledged “lapses in safeguards,” stating that fixes were being implemented “urgently.” This admission came alongside statements from xAI employees indicating efforts to “tighten” the chatbot’s protective measures, but the damage – and the questions – remain.
The very nature of Grok’s integration into X allows for direct interaction; users can tag the chatbot in posts and request responses, resulting in generated text and images appearing as regular posts on the network. This accessibility, while intended to be engaging, has clearly been exploited.
This incident underscores a growing crisis in the realm of AI-generated imagery. The ability to create realistic depictions of undressed minors presents immense challenges for content moderation and the safety systems built into large language models. Existing “guardrails” are proving vulnerable to manipulation.
The Internet Watch Foundation reported a staggering 400% increase in AI-generated child sexual abuse imagery in the first half of 2025 alone. This exponential rise highlights the urgent need for rigorous testing and robust safeguards before such powerful tools are released to the public.
While xAI has positioned Grok as more lenient than other AI models, even introducing a “Spicy Mode” allowing for some adult content, the company explicitly prohibits pornography involving real people and, crucially, any sexual content involving minors. The recent events demonstrate a clear failure to uphold these stated principles.
Reports indicate users successfully prompted Grok to digitally alter images, removing clothing from subjects – predominantly women – leaving them in minimal attire. This prompted a formal review request from India’s IT ministry, demanding a comprehensive assessment of Grok’s safety features.
The French government has already reported the illicit content to its public prosecutor, seeking its immediate removal. This swift action reflects the gravity of the situation and the determination to hold those responsible accountable.
Other leading AI companies, like OpenAI and Google, have implemented policies prohibiting the sexualization of minors and banning users who attempt to generate such material. However, the discovery of child sexual abuse material within the datasets used to train these AI models reveals a systemic problem.
In 2023, researchers uncovered over a thousand instances of child sexual abuse material within a massive public dataset used to build popular AI image-generators. This underscores the difficulty of ensuring clean data and preventing the replication of harmful content.
This isn’t an isolated incident. Meta Platforms Inc. recently faced criticism for policies that allowed its chatbot to engage in inappropriate conversations with children, prompting a policy update. The pattern reveals a consistent struggle to protect vulnerable individuals in the age of increasingly sophisticated AI.
