Data shows hundreds of non-consensual AI images were created by Grok on X

A new study surveying X users who approach Elon Musk’s AI chatbot Grok shows how often people use it to create sexualized images. Almost three-quarters of the posts collected and analyzed by a doctoral student at Trinity College, Dublin, were requests for non-consensual images of real women or minors having clothing removed or added.

The posts provide a new level of detail into how the images are generated and shared on X, with users coaching each other using prompts; Suggestions for repeats of Grok’s depictions of women in lingerie or swimsuits or with body areas covered in semen; and in replies to posts containing self-portraits from female users, Grok was asked to remove his outer clothing.

Among hundreds of posts that Nana Nwachukwu identified as direct, non-consensual requests for Grok to remove or replace clothing, dozens reviewed by the Guardian show users posting images of women, including celebrities, models, stock photos and women who are not public figures…