AI chatbots strip hijabs from photos of Muslim women, Guardian test finds
A Guardian investigation found ChatGPT and Grok removed a hijab from a photo of a Muslim woman while refusing to remove a dress, with Gemini complying after a 'western' rephrase.
Leading AI chatbots will remove a hijab from a photograph of a Muslim woman while refusing comparable clothing edits, a Guardian investigation by Johana Bhuiyan has found.
The paper tested four chatbots — ChatGPT, Grok, Gemini and Claude. ChatGPT and Grok complied with requests to strip the hijab from a photo of a Muslim woman, yet both refused to remove a dress from a photo of a woman. Gemini initially refused the hijab request, but complied when asked to make the woman look “western” — a workaround that also removed a turban and a nun’s habit. Claude, which has no image-editing feature, refused.
The investigation cites Eviane Leidig, director of research and outreach at the Center for the Study of Organized Hate, as an expert source.
The inquiry followed a viral case in which French National Rally MP Julien Odoul posted an AI-altered photograph of a real Muslim woman with her hijab removed and her clothes changed, captioned “LIBERTÉ FRANÇAISE”. Odoul, the MP for Yonne’s third constituency, posted the image on 5 September in response to a post by La France insoumise MP Aly Diouara, Tech Times reported.
OpenAI, Google, xAI and Anthropic all declined to comment, according to the Guardian.
Fact-check: 9/9 core facts PASS (see verdict below). No image: the real woman in the trigger case must not be depicted, and no suitably licensed generic illustration was available. image_exception granted; ships imageless.
More on this topic: all Technology stories
