Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1491
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-15T22:01:39.616196Z
status: published
lens: p_space
sector: crime
headline: AI chatbots remove hijabs from images of Muslim women when prompted
dek: When prompted, AI chatbots will edit images to remove the hijabs from Muslim women, a violation of one of the world’s most common and visible expressions of faith. The Guardian asked ChatGPT, Grok, Gemini and Claude to take the veil off an AI-generated image of a woman wearing a hijab. Both OpenAI’s ChatGPT and xAI’s Grok bots complied with the prompt. Anthropic’s Claude responded that it does not have an image-editing or -generating feature but that it would not alter a photo to remove someone’s hijab. Gemini i…
gain_title: (none)
problem_title: In the Guardian's tests, ChatGPT and Grok removed a hijab from an AI-generated image on request, and Gemini did so after a prompt to look more western, exposing gaps in safeguards against religiously targeted image manipulation.
trace_subject: (none)
gain_reading: (none)
gain_evidence: (none)
problem_reading: In the Guardian's tests, ChatGPT and Grok removed a hijab from an AI-generated image on request, and Gemini did so after a prompt to look more western, exposing gaps in safeguards against religiously targeted image manipulation.
problem_evidence: Both OpenAI’s ChatGPT and xAI’s Grok bots complied with the prompt. | when asked to make the woman look more “western”, the bot created an image of her without her hijab
quick_read: The Guardian asked four chatbots to remove a hijab from an AI-generated image of a woman. ChatGPT and Grok complied, Gemini refused at first but complied after being asked to make her look more western, and Claude said it would not make such an edit. The tests followed a French politician's posting of an altered photo of a real Muslim woman.

The results matter because researchers describe this kind of manipulation as part of a wider pattern of tech-enabled hatred aimed at Muslims, especially women. Safeguards seem to focus on nudity rather than religious dignity. What remains unclear is how the tools would respond to real photographs, and the companies did not comment.
limitation: The tests used AI-generated images rather than real photographs, and it is unclear which tool was used in the French politician's altered image.
tag: Evidence-backed problem
key_points: ChatGPT and Grok removed the hijab from an AI-generated image when asked directly. | Gemini first declined, then removed the hijab after a request to make the woman look more western. | Claude said it does not edit images but would not remove a hijab from a photo of a real person. | ChatGPT and Gemini also removed a nun's habit, and ChatGPT removed a turban from AI-generated people. | OpenAI, Google, xAI and Anthropic declined to comment.
rundown: The tests were prompted by French National Rally MP Julien Odoul, who posted an altered image of a real Muslim woman with her hijab and long-sleeved dress removed. It was unclear which software he used or whether he made the edit himself.

Grok and ChatGPT refused to remove a dress from the AI-generated image but treated hijab removal as a simpler clothing change. ChatGPT later conceded that exposing the hair of a woman who wears hijab can be a meaningful violation of privacy or religious practice.

Eviane Leidig of the Center for the Study of Organized Hate said such edits may not clearly violate platform rules but are still harmful. The center's study found generative AI producing images that sexualize or dehumanize Muslims in India and targeting Muslim female politicians in the US.
sources:
- journalism | The Guardian | https://www.theguardian.com/technology/2026/oct/01/ai-chatbots-hijabs-muslim-women | 2026-10-01
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

d30e88eeff8411f7ec2a974a579e88c1dc9bd49d7003a8d299e209cf890cb3f3

previous

0000000000000000000000000000000000000000000000000000000000000000