In August last year, Arve Hjalmar Holmen, a Norwegian man, asked ChatGPT a question out of boredom. He typed in his name and asked, “Who is Arve Hjalmar Holmen?” The response he received out of simple curiosity came as a shock.
That’s because ChatGPT identified him as a murderer who killed his two sons.
On March 21, BBC reported that Holmen had filed a defamation lawsuit against OpenAI, the company behind ChatGPT, for providing false information.
The AI-generated response claimed that Holmen was the father of two boys, aged 7 and 10, who were found dead near a pond by their home in Trondheim, Norway, in December 2020. It added that he had been sentenced to 21 years in prison, the maximum penalty in Norway, for attempting to murder his third son as well.
NOYB, the digital rights organization that filed the complaint on Holmen’s behalf, stated that the response was defamatory and violated the European Union’s General Data Protection Regulation (GDPR), which requires the accuracy of personal data. The group emphasized that Holmen is a law-abiding citizen who has never been charged with or convicted of any crime.
The Holmen filed the complaint because ChatGPT’s information was accurate, while other parts contained critical errors.
Details such as the number of children and their ages matched Holmen’s accurate information. According to the BBC, Holmen believed these elements could cause serious harm to reputation.
Eventually, through NOYB, he filed a complaint with Norway’s Data Protection Authority, requesting that OpenAI be fined.
ChatGPT displays a notice at the bottom of the screen stating, “ChatGPT can make mistakes. Check important info.” Holmen’s side criticized this message as a mere attempt to avoid responsibility.
Holmen’s attorney, Joakim Söderberg, argued that spreading false information and simply adding a small disclaimer at the end stating that everything said may not be true is unacceptable. He emphasized that personal information must be accurate and that individuals must have the right to correct it.
OpenAI has not yet issued an official response.
