Eric Bothorel MP, from the Renaissance group, announces that he has filed a complaint this April 12 with the National Commission for Informatics and Liberties (Cnil) due to the inaccuracy of the information generated by ChatGPT about him.
Although the first two French complaints against ChatGPT were filed at the beginning of April with the Cnil, the parliamentarian explained to Tech&Co that he made this referral under article 5.1.d of the GDPR, the general data protection regulation, which requires that personal data are “accurate”. and, where appropriate, updated” and that inaccuracies “are deleted or rectified without delay” using “all reasonable measures”.
The chosen one, particularly focused on digital issues, says he sent several requests for his bio on the free-access version of the chatbot, and each time was faced with false information.
“The constituency [des Côtes d’Armor où il est député] it is not the correct one. I have never been mayor of any town. [ChatGPT indique qu’il l’a été à Lannion et Saint-Brieuc]. I have never been a regional councillor. I have never worked at Havas or Orange”, enumerates Eric Bothorel, who also jokes about the fact that the chatbot has made him younger by giving him the wrong date of birth.
A fallible tool
This complaint once again brings to light the “hallucinations” of ChatGPT, that is, information manufactured by the AI. In fact, this type of generative AI sorts words based on how likely they are to be linked together, producing many erroneous results.
In early March, an Australian mayor threatened to file a libel suit against OpenAI, the startup behind ChatGPT, because the chatbot falsely claimed it had served jail time for corruption.
To enforce this article 5.1.d, Eric Bothorel asks this type of AI to “respect the provisions written in the GDPR, […] that is to say, to make the necessary efforts so that the erroneous information ceases to be erroneous”.
Source: BFM TV
