ChatGPT Falsely Accuses Man of Murdering His Children
The man in question filed a complaint against OpenAI due to a disturbing incident involving the AI system. While Generative AI is known for creating "hallucinations," or false information in its responses, this particular instance crossed a serious line.
The Disturbing Encounter
An individual named Arve Hjalmar Holmen decided to test what ChatGPT, an AI language model, knew about him. To his shock, the AI claimed that he had murdered his own two sons and had received a 21-year prison sentence. What made this claim even more chilling was that the AI included specific details from his personal life, such as the number and gender of his children, and even his town of residence.
As noted by the privacy rights advocacy group Noyb, this blatant misinformation violated data protection laws, specifically the GDPR. Joakim Söderberg, a data protection lawyer at Noyb, emphasized that personal data must be accurate, and users have the right to request corrections if necessary.
Legal Actions Taken
Shocked and disturbed by the false accusations, Holmen filed a formal complaint against OpenAI, urging the company to remove the erroneous output, refine its AI model, and limit the processing of his personal information. In response, OpenAI took corrective measures to prevent such incidents from recurring.
Continued Scrutiny and Legal Processes
Noyb pursued the matter further by filing another complaint against OpenAI in April 2024, adding to the pressure for accountability. However, as of the latest update, the case remains unresolved, with legal proceedings ongoing.
As the investigation continues, it raises broader concerns about the responsibility of AI companies to comply with data protection regulations and prevent the dissemination of false information that can harm individuals' reputations.
