A Norwegian man has filed a complaint after ChatGPT generated false information, claiming he murdered his two sons and was sentenced to 21 years in prison.
Arve Hjalmar Holmen reported the issue to the Norwegian Data Protection Authority, demanding OpenAI face penalties for the damaging error.
His case highlights ongoing concerns about AI-generated misinformation, known as “hallucinations,” where artificial intelligence produces and presents false details as fact.
Holmen discovered the error when he asked ChatGPT about himself.
The AI responded with a fabricated story, stating that he was responsible for the deaths of his two sons, whose ages it estimated correctly, and that he had attempted to kill a third child. He described the situation as alarming, fearing that people might believe the claims.
Digital rights organization Noyb, which is backing his complaint, argues that OpenAI violated data protection rules requiring accuracy in personal data.
The group criticized OpenAI’s disclaimer stating ChatGPT “can make mistakes,” arguing that it does not justify the spread of harmful falsehoods.
OpenAI responded that the incorrect response came from an older version of ChatGPT and that its latest updates, which include real-time search, improve accuracy.
AI hallucinations remain a significant issue, with other AI systems also generating false information. Apple recently suspended its AI-powered news summaries in the UK after errors, and Google’s Gemini chatbot has also been criticized for inaccurate claims.
Experts say the cause of these AI-generated fabrications is still not fully understood, even by those developing the technology. Holmen’s complaint raises broader concerns about accountability and the risks of misinformation created by artificial intelligence.
