xAI acknowledged on Friday that lapses in its safety systems led to the generation of “images depicting minors in minimal clothing” on the social media platform X, following user interactions with its AI chatbot Grok.
According to statements posted on X, screenshots shared by users showed Grok’s public media tab populated with altered images after users uploaded photos and prompted the bot to modify them. In response, Grok said there were “isolated cases” in which users successfully obtained such images, despite existing safeguards designed to prevent harmful outputs.
The chatbot emphasized that while xAI has protections in place, the company is “urgently fixing” identified weaknesses to fully block requests involving minors. Grok explicitly stated that CSAM—child sexual abuse material—is illegal and strictly prohibited, underscoring that preventing such content is a priority.
In a separate reply to a user the previous day, Grok noted that most incidents could be avoided through more advanced filters and monitoring systems. However, it added a familiar caveat in AI safety discussions: no automated system is entirely foolproof. xAI said it is reviewing user-reported cases and accelerating improvements to close remaining gaps.
When contacted for comment by Reuters, xAI responded curtly via email with the phrase “Legacy Media Lies,” offering no additional clarification. The incident adds to broader scrutiny of generative AI tools and their ability—or failure—to enforce strict content boundaries, particularly on platforms associated with Elon Musk.
