Privacy advocacy group Noyb filed a formal complaint with Norway’s data protection authority against OpenAI, alleging the company violated EU data laws by failing to correct ChatGPT’s fabrication of a “defamatory” story falsely implicating a Norwegian man in a fictional child murder case.
The Allegations
The complaint centers on a Norwegian user, Arve Hjalmar Holmen, who discovered that ChatGPT had generated a detailed, entirely fictitious narrative alleging he was convicted of murdering two of his children, attempting to kill a third, and serving a 21-year prison sentence. The AI’s output shockingly wove real personal details—including his hometown, the number and genders of his children—into the false claims, amplifying risks to his reputation.
Noyb argues that OpenAI’s failure to provide users with mechanisms to correct such errors breaches Article 5(1)(d) of the GDPR, which mandates that personal data be “accurate and, where necessary, kept up to date.” The group emphasizes that ChatGPT’s tendency to “hallucinate” damaging falsehoods—ranging from bribery and abuse accusations to murder—has repeatedly exposed individuals to harm, with no clear recourse for correction.
AI Hallucinations: From Quirks to Legal Risks
ChatGPT’s “hallucinations,” a term describing its generation of plausible but fabricated information, have drawn scrutiny since the chatbot’s launch. While some inaccuracies are harmless, cases like Holmen’s underscore their potential severity. Media reports and lawsuits have previously highlighted instances where the tool falsely accused individuals of crimes, including embezzlement and sexual misconduct.
OpenAI has responded by adding disclaimers noting ChatGPT may produce errors. However, Noyb’s data protection lawyer, Joakim Söderberg, dismissed this as insufficient: “The GDPR is clear. You can’t spread false information and hide behind a tiny disclaimer. Users have the right to correct inaccurate data—a principle OpenAI systematically ignores.”
Implications
The case highlights a critical clash between rapidly evolving AI technology and established data rights. With regulators increasingly scrutinizing AI systems, the outcome could set a precedent for how generative AI tools handle accuracy, accountability, and user protections under EU law. For now, victims of ChatGPT’s “creative” storytelling wait to see if the law will compel OpenAI to rewrite its approach to the truth.
