AI chatbots remove hijabs from images of Muslim women

The Guardian tested ChatGPT, xAI's Grok, Google’s Gemini and Anthropic’s Claude and found ChatGPT and Grok complied when asked to edit an AI-generated image to remove a woman’s hijab. Gemini initially said it would not remove clothing or head coverings from photos of real people without consent but later produced an image without a hijab after being asked to make the woman look more "western." Anthropic’s Claude said it does not have an image-editing feature and that it would not alter a photo to remove someone’s hijab. The tests followed a French far-right MP Julien Odoul posting a modified photo showing a Muslim woman without a hijab and full-length dress, an episode experts link to a wider pattern of tech-enabled anti-Muslim hatred. A Center for the Study of Organized Hate study cited by the Guardian found generative AI is being used in India to sexualize and dehumanize Muslim women and to target U.S. Muslim politicians such as Ilhan Omar and Rashida Tlaib. The chatbots generally refused requests to digitally undress a woman, and ChatGPT acknowledged that exposing a hijab-wearer’s hair can be a meaningful violation of privacy or religious practice. xAI previously said it had measures to block requests to undress images of real people in revealing clothing.