Summary of the Complaint and Response
The Norwegian man, Arve Hjalmar Holmen, has filed a formal complaint with the Norwegian Data Protection Authority, claiming that an AI-powered chatbot named ChatGPT has falsely accused him of murdering his two sons. The complaint was filed in December 2020, and the man described himself as a “regular person” with no public or official record, focusing solely on a tragic event involving the innocent children of a Norwegian father. The AI bot was asked about his identity, and it provided a tentative response accusing him of the玛ful events. This response was problematic, as it included details such as the number of his children, their age gap, and his home location, which could have significant implications for privacy.
The company behind ChatGPT, OpenAI, received widespread attention upon the woman’s complaint because it was associated with a well-known delivery service in Norway where the mailman was claimed to have killed his children. The response from ChatGPT began with a repeating of the man’s name and locale, which was immediately deemed false, and claims he killed both of his children age 14 and 10 in a pond. The response also included the_man’s full address and contact information, further inflating the false narrative.
The Norwegian Data Protection Authority has filed a complaint with the Norwegian Data protection authority for the AI bot’s response. ThemannDe wrote in the complaint that the false outputs are “completely false,” but that the facts are still very much “in the same ballpark” as the man’s own life. Theman pointed out that the complaint allegedly violated the GDPR’s markdown requirements. The complaint also criticized the paragraph for having “no support for the case against the man’s official legal or medical institutions.” The complaint has now been filed by a digital rights group, Noyb, which charges the AI bot as having been “never accused of any crime, or ever guiltless in any regard,” to undermine the “defamatory” outputs that were deemed false.
In a separate statement, OpenAI welcomed the complaint and defended its AI response. The company stated that the model it released in 2017 continues to evolve, and that some aspects of its responses are more likely to contain factual errors. The new model developed from 2018 onward, which is more extensive, now includes features like online search capability, and the company claims that it would reduce claims of hallucinations and promotion of misinformation. OpenAI further mentioned that the email received from Holmen was “untrue, but in the message of the usual accuracies that we rely on.”
The department allows for repetition errors in some of its outputs, unlike other AI models. Its sentences are generated using an autotuning transcript, which explains why in the complaint the AI bot repeated the man’s address, but the wing of his address is entirely different. OpenAI emphasized that the bot’s more recent AI would reduce the chance of such inaccurate outputs.
The legal expert mentioned that the complaint includes elements similar to the man’s actual life, such as his home location and children’s details. The expert also indicated that the only thing that could exacerbate the problem is if the output was — and is —$copied$ from the original. The expert askedOpenAI’s parent to modify its responses to include accuracy and prevent false claims that are unverified. Finally, Noyb mentioned that since the AI bot responded to the woman’s complaint, OpenAI had released a new model with specific search features, reducing the chance of similar errors from now on.
The AI bot, which uses a “fewer” working model, spent three years offsetting the original’s super-slow response time, resulting in a more accurate acknowledgment of the facts. The conversation between the man and the bot is still a mess, but for ten years would no longer exist if the woman were to file this complaint againstOpenAI. The AA bot, while ahighest-known AI, continues to be a reliable tool for generating factual answers and accurate reports. The company, with its strengths in creativity and regulation — which it brought to the table opposite the