Getty Images
The Internet Watch Foundation (IWF) charity states its analysts found “criminal imagery” of girls, aged 11 to 13, which seems to have been generated by Grok.
Grok, an AI tool owned by Elon Musk’s xAI firm, is accessible via its website, app, and the social media platform X.
The IWF reported discovering “sexualised and topless imagery of girls” on a dark web forum, where users asserted they utilized Grok to produce the content.
X and xAI have been contacted for comment.
Ngaire Alexander of the IWF expressed concern that AI tools like Grok risk “bringing sexual AI imagery of children into the mainstream.”
She noted that under UK law, such material would be classified as Category C, representing the lowest severity of criminal content.
However, she added that the uploader subsequently used a different AI tool, unrelated to xAI, to generate a Category A image, which is the most severe classification.
“There is extreme concern regarding the apparent ease and speed with which individuals can generate photo-realistic child sexual abuse material (CSAM),” she stated.
The charity, whose mission is to remove child sexual abuse material from the internet, runs a hotline for reporting suspected CSAM and employs analysts to evaluate the material’s legality and severity.
IWF analysts located the material on the dark web; these images were not found on the social media platform X.
Ofcom had previously contacted X and xAI after receiving reports that Grok could be used to create “sexualised images of children” and digitally undress women.
Examples have been observed on the social media platform X where individuals requested the chatbot to modify real images, depicting women in bikinis without consent or placing them in sexual scenarios.
The IWF confirmed receiving reports of similar images on X, though these had not yet been determined to meet the legal definition of CSAM.
In a prior statement, X indicated its policy: “Action is taken against illegal content on X, including CSAM, through removal, permanent account suspension, and cooperation with local governments and law enforcement as required.
“Any user prompting Grok to generate illegal content will face the same repercussions as if they uploaded illegal content directly.”
