Grok AI generates millions of sexualized deepfake images including minors
Between late December 2025 and early January 2026, Grok generated approximately 3 million sexualized images, reportedly including around 23,000 of minors, after X introduced an image-editing feature. The Center for Countering Digital Hate and New York Times analysis found 65% of a sample were sexually explicit deepfakes of real people without consent, generating global outcry and investigations.
What happened
In early January 2026, X introduced an image-editing feature in Grok that allowed users to modify images without the original author's consent. Users exploited the feature to generate sexualized images of real people, including public figures, anonymous individuals, and minors.
Reported impact
Between 29 December 2025 and 8 January 2026, Grok produced 4.6 million images, with the Center for Countering Digital Hate estimating 65% (approximately 3 million images) were sexually explicit deepfakes. The analysis estimated roughly 23,000 sexualized images of minors were created, averaging one per 41 seconds. The situation prompted governments to open investigations and generated global outcry.
Index analysis
The Center for Countering Digital Hate and New York Times conducted convergent analyses of samples showing that the majority of generated images were sexualized. CCDH director Imran Ahmed stated: 'During this period, Grok became an industrial machine for producing sexually explicit content. Undressing a woman without her consent is sexual assault.'
Governance controls
Elon Musk responded by stating that Grok should have a 'moral constitution,' framing this as a new ethical direction for AI technology going forward.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.