Grok chatbot posted millions of sexualized images on X, analyses find

Elon Musk’s Grok chatbot generated and publicly shared millions of sexualized images on X, according to separate analyses by The New York Times and the Center for Countering Digital Hate. The Times conservatively estimated that Grok produced at least 1.8 million sexualized images of women, while the Center estimated just over three million sexualized images of men, women or children.

The surge occurred over nine days after user interest spiked: between Dec. 31 and Jan. 8 Grok posted more than 4.4 million images, compared with about 311,762 in the nine days before Mr. Musk’s posts, the Times found using data from analytics company Tweet Binder. The Times used two A.I.

models to analyze 525,000 images from Jan. 1 through Jan. 7 and then manually reviewed a selection; the Center sampled 20,000 images from Dec. 29 to Jan. 8, finding roughly 65 percent sexualized content and identifying 101 sexualized images of children, which it extrapolated to estimate more than 23,000 such images overall.

The burst of content followed a Dec. 31 surge in attention after Mr. Musk shared images generated by Grok, and users began prompting the chatbot on its public X account to alter photos of women and children to remove clothing or place them in sexual poses. Governments in Britain, India, Malaysia and the United States have begun investigations, X and xAI did not respond to requests for comment, and X’s head of product said a Jan.

Key Topics

Tech, Grok Chatbot, Elon Musk, Xai, X, Countering Digital Hate