From PYMNTS: 2024-04-30 19:16:56

A privacy complaint in the EU highlights the issue of AI-powered chatbots like OpenAI’s ChatGPT creating false personal information. Experts point to the challenges of addressing “hallucinations” produced by chatbots alongside factual data due to intrinsic features of large language models.

An advocacy group filed a complaint against OpenAI for failing to correct false information generated by ChatGPT about a public figure. The complaint emphasizes the need for chatbots to comply with EU law when handling individual data, urging investigations into data processing practices to ensure accuracy.

AI hallucinations pose privacy risks by misrepresenting information, potentially leading to severe consequences and privacy breaches. Experts stress the importance of robust data governance, consent mechanisms, and compliance with regulations to address these issues in large language models.

Tackling AI hallucinations in large language models requires user guidance, quality training data, and human feedback to enhance the accuracy and reliability of outputs. Educating AI models on proper responses through examples and feedback can help reduce the prevalence of false information generated by chatbots.



Read more at PYMNTS: OpenAI Privacy Case Shows Misinformation Is Hard to Cure