Osegi
AI·2 min read

Grok lawsuits: ordinary school photos turned into abuse images

A woman learned from an investigator that her photo from age 14 was fed to Grok. She is now one of a growing group of plaintiffs suing xAI.

A teenager's school graduation photo on a phone screen, with a chatbot interface reflected in the glass.
Image: Spider's Web
SUMODialoediviewergiftdetect

A graduation photo from primary school, the kind that sits in a family album. According to a lawsuit, that was all it took. A woman, now 18, says an investigator told her that an adult had used Grok, the chatbot from Elon Musk's xAI, to turn a picture of her at 14 into sexualised abuse material. She has joined a suit against xAI.

What the Grok lawsuits claim

The photo showed a group of pupils at a school ceremony. Nothing intimate. The woman says she hadn't thought about it for years, until a law enforcement officer showed it to her and explained what had been done. The fake was then spread online, and the complaint notes there is no way to know how many people have it.

She is not alone. In March, minors in Tennessee sued xAI, claiming Grok was used to sexualise their ordinary photos, and the case later gained more plaintiffs. Parents and minors in Arkansas filed similar claims. In August, a nationwide class action followed in California. The plaintiff there also alleges that abuse material from her childhood was in the data used to build the system, and that Grok then generated more content with her likeness. Lawyers estimate thousands of people could be affected. These are allegations in court filings, not findings.

Analyses cited by Bloomberg say that at one point Grok was producing thousands of sexualised or "undressing" images an hour on X. Others suggested tens of thousands of realistic sexualised images of children could be among the results.

Is this only a US problem?

No. Poland's research institute NASK reported in May that AI-generated child abuse material found by its Dyżurnet.pl hotline rose 300 percent in a year, from 88 cases in 2024 to 250 in 2025. The hotline points to "nudify" apps and to material built from real children's faces. Its head, Martyna Różycka, warns that a photo a parent posts can later be used in a fake sexual image.

Polish lawmakers have reacted. A bill sent to the Sejm on 7 October would criminalise sharing digitally created or altered nude or sexual images of a person without consent, explicitly including AI. The proposed penalty is 3 months to 5 years in prison, the same as for real images. In the EU, AI Act transparency rules for AI-generated content, including deepfakes, have applied since 2 August.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:grokxaielon muskdeepfakelawsuitchild safety

More from this day

See all →

More in: AI

See all →