ChatGPT and Grok remove hijabs from photos on request

A review of Grok’s activity between January 6–9, 2026, showed about 5% of image edits involved altering religious clothing. One viral example saw a user with 180,000+ followers prompt Grok to modify images of women into revealing outfits, which were viewed over 700,000 times. Beyond explicit requests, some AI image generators have been found to strip hijabs and modest clothing without user prompts, though xAI dismissed these claims with automated replies. The incidents highlight a regulatory gap: while deepfakes and election misinformation dominate AI safety debates, non-consensual edits targeting religious attire remain unaddressed. Neither company has publicly addressed whether the individuals in edited photos consented to the changes.