The European data protection advocacy group noyb has complained to OpenAI, alleging violations of the General Data Protection Regulation (GDPR) due to inaccuracies in data generated by ChatGPT. According to noyb, OpenAI’s failure to rectify erroneous personal data breaches GDPR standards.
Maartje de Graaf, Data Protection Lawyer at noyb, emphasized the severity of inaccuracies, stating that false information about individuals can lead to serious consequences. She asserts that technology must adhere to legal requirements, not the other way around.
Under the GDPR, personal data must be accurate, with individuals having the right to rectification and access to information about their data. However, OpenAI has admitted its inability to correct ChatGPT’s inaccuracies or disclose data sources.
What does OpenAI have to say?
OpenAI argues that factual accuracy in large language models is an ongoing research area. However, noyb points to a New York Times report revealing that chatbots like ChatGPT invent information up to 27% of the time. In one instance cited by noyb, ChatGPT repeatedly provided an incorrect date of birth despite rectification requests.
Despite OpenAI’s assertion that it could filter or block certain data, it failed to adequately respond to the complainant’s access request, violating GDPR obligations. The complaint seeks an investigation by the Austrian Data Protection Authority and corrective measures to ensure compliance with GDPR standards.
