Elon Musk’s xAI under fire as Grok chatbot faces scrutiny over explicit and disturbing content

Grok AI chatbot, which is built by Elon Musk’s artificial intelligence venture, xAI, is again under scrutiny after a new report reveals that it allows users to access uncensored content on the platform. This has raised alarms about safety and the potential generation of unlawful content on the chatbot, especially when it comes to child safety.

According to Business Insider, out of 30 current and past xAI staff training Grok, at least 12 have particularly highlighted its controversial modes, labelled “sexy” and “unhinged”, as being exposed to large volumes of explicit content during training, including requests to generate AI-based child sexual abuse material (CSAM). The report also mentions that Grok AI has a flirtatious female avatar, which is capable of ‘undressing on command,’ raising alarms about the chatbot’s potential misuse for children.

Further, these employees have reportedly revealed that they were required to assess hundreds of not-safe-for-work(NSFW) images, videos, and audio files as part of training the chatbot. Notably, the risks escalated after xAI rolled out ‘Project Rabbit,’ which is aimed at improving the conversational skills of GrokAI. Workers claimed they were tasked with transcribing user chats, where one worker reportedly said the role involved listening to ‘audio porn,’ describing the experience as ‘eavesdropping’ as users were unaware that their interactions might be reviewed.

This is not the first time Grok AI has faced criticism over its content suggestions. Earlier in 2025, Grok Imagine was found generating deepfake-style images of Taylor Swift, some in sexually explicit outfits, whereas previously the chatbot praised Adolf Hitler, even referring to itself as 'MechaHitler'.