AI chatbots can remove hijabs from images of Muslim women, Guardian investigation finds

AI chatbots can remove hijabs from images of Muslim women, Guardian investigation finds
—————————————–
Some generative AI tools can be prompted to digitally remove the hijab from images of Muslim women, raising concerns over consent, privacy and the potential misuse of image-generation technology.

An investigation by The Guardian found that ChatGPT and Grok complied with requests to remove a Muslim woman’s hijab from an image. Google’s Gemini initially refused the request but generated an unveiled version after the instruction was reframed as making the woman appear “more Western”. Claude, meanwhile, refused to carry out the request.

The findings emerged amid controversy in France after Julien Odoul, a member of the right-wing National Rally party and a French MP, shared an altered image of a Muslim woman on social media. The edited version showed the woman without her hijab and long clothing. It remains unclear which software or AI tool was used to alter the image.

The issue has also been highlighted by research from the Center for the Study of Organized Hate (CSOH), which has documented the use of AI-generated imagery to spread anti-Muslim content. Its research in India examined more than 1,300 AI-generated images and videos targeting Muslims, including material sexualising Muslim women, promoting derogatory stereotypes and depicting violence.

Researchers have also raised concerns about AI-generated content targeting prominent Muslim women and politicians in the United States.