Imagine asking ChatGPT what he knows about you, and finding out he's accusing you of murdering your own children! This is the nightmare experienced by a Norwegian father.
The artificial intelligence ChatGPT recently distinguished itself by accusing wrongly accused a Norwegian father of infanticide. A mistake that raises serious questions about the reliability of chatbots and their potential consequences.
Indeed, when Arve Hjalmar Holmen, a Norwegian man with no prior history, asked ChatGPT what it knew about him, he was far from imagining the answer he would receive. The chatbot described him as a murderer who had killed two of his children and attempted to murder the third, all accompanied by a 21-year prison sentence. A completely unfounded accusation that deeply shocked this man.
Read also – ChatGPT is coming to your Google Drive and Slack, a revolution for teleworking?
A father unfairly accused by AI
The most disturbing thing about this case is that ChatGPT mixed true information with its false accusations. The number of children, their gender, and even Arve Hjalmar Holmen's hometown were correct. This combination of truth and falsehood makes the error all the more credible and dangerous.
Faced with this excess, the Austrian privacy association Noyb has stepped up to the plate. She filed a complaint against OpenAI, the company behind ChatGPT, for violating the General Data Protection Regulation (GDPR). According to Noyb, a simple warning about the chatbot's potential errors is not enough to absolve OpenAI of liability.
This is not the first time that ChatGPT has been criticized for its "hallucinations," the term used to describe these inventions presented as fact. Other cases have been reported, involving false accusations of fraud, child abuse, and sexual harassment.
The case raises serious questions about the use and regulation of AI chatbots. How can we ensure the accuracy of the information they provide? What are the responsibilities of the companies that develop them? The debate is far from over and is likely to intensify as these technologies become more widespread in our daily lives. In the meantime, Arve Hjalmar Holmen hopes to obtain redress and, above all, that such errors will not be repeated in the future.
0 Comments