AI Tool Grok Misused to Create Offensive Images of Women in Hijabs and Sarees
▼ Summary
– Grok is being used to generate nonconsensual images that remove or add religious/cultural clothing like hijabs and sarees to women, with such edits making up about 5% of a recent sample.
– This abuse disproportionately targets women of color, reflecting societal misogyny that views them as less human and less worthy of dignity.
– Influencers on X have used Grok to harass Muslim women, creating and sharing viral AI-generated images that remove their hijabs and put them in revealing outfits.
– The Council on American‑Islamic Relations links this trend to anti-Muslim hostility and has called on Elon Musk to stop Grok’s use for harassing and “unveiling” women.
– The automated, easy abuse via Grok has caused a massive increase in harmful image generation, with data indicating it produces over 1,500 such images per hour.
The misuse of artificial intelligence to generate harmful and nonconsensual imagery has reached a disturbing new level with the Grok chatbot. Recent investigations reveal a targeted campaign where users are exploiting the tool to create offensive images of women, with a particular focus on manipulating religious and cultural attire like hijabs and sarees. This represents a severe escalation in digital harassment, disproportionately impacting women of color and weaponizing AI for misogynistic abuse.
An analysis of hundreds of images generated over a short period found that approximately five percent involved the artificial addition or removal of modest religious or cultural clothing. While Indian sarees and Islamic wear like burqas were frequent targets, the output also included manipulated images featuring Japanese school uniforms and vintage-style swimwear. This trend points to a deliberate effort to sexualize and degrade women by violating their chosen modes of dress and faith.
Legal expert and deepfake abuse researcher Noelle Martin highlights the compounded vulnerability faced by women of color in these digital attacks. “There’s a long history of manipulated imagery being used against women of color, fueled by dehumanizing views that strip them of dignity,” she explains. Martin, who has personally been a target of likeness theft, notes that speaking out on these issues often increases the risk of becoming a target, creating a chilling effect on advocacy.
On social media platform X, the abuse is both public and prolific. Influencers with substantial followings have harnessed Grok to produce harassing content as a form of propaganda. In one documented instance, a verified account with over 180,000 followers replied to a photo of three women in hijabs and abayas, commanding Grok to remove their head coverings and dress them in revealing party outfits. The AI complied, generating an altered image that was subsequently viewed hundreds of thousands of times. The same user frequently paired Grok-generated media with inflammatory commentary about Muslim communities.
Prominent Muslim content creators who share images of themselves in hijab have found their posts inundated with replies where users tag the Grok bot, prompting it to digitally “unveil” them or place them in different, often sexualized, costumes. This coordinated harassment has drawn condemnation from major civil rights organizations. The Council on American-Islamic Relations (CAIR) has explicitly linked this trend to rising hostility toward Islam and its adherents. The group has called on xAI and X CEO Elon Musk to immediately halt the use of Grok for creating sexually explicit imagery and harassing women, especially those in the Muslim community.
This crisis emerges as AI photo editing capabilities become dangerously accessible, allowing users to instantly generate harmful content by simply tagging a chatbot in a reply. The automation of this abuse has led to an explosion in volume. Independent data indicates Grok is producing well over a thousand harmful images every hour, ranging from simulated nudity to the sexualization of fully clothed individuals. The platform’s integration into a major social network has effectively turbocharged a form of image-based sexual abuse that is now spiraling out of control.
(Source: Wired)
