Will French law soon be enriched by artificial intelligence? The bill relating to the 2024 Olympic and Paralympic Games is currently being debated in the National Assembly, which includes in particular the possibility of surveillance using “smart” cameras, capable of detecting abnormal behaviour. It was recently enriched with an amendment that was original to say the least: it was written by the artificial intelligence ChatGPT.
Co-signed by about twenty deputies from the Libertés, Indépendants, Outre-mer et Territoires group, at the initiative of Jean-Félix Acquaviva deputy (2nd district of Upper Corsica), the amendment was created by software designed by OpenAI company, capable of generate complex texts at the request of any Internet user. A deliberately provocative method, in order to warn of the dangers of technology, if used in the context of mass surveillance.
retouch in the margin
Identified by Letter A, the amendment proposes specifying in the text that the analysis of data by video surveillance cameras must be carried out with “respect for the fundamental rights and freedoms of the persons concerned and the obligation to guarantee humane decision-making in each stage of processing.
With Tech&Co, Jean-Félix Acquaviva specifies that he used the previous version of ChatGPT (now superseded by GPT-4) to create the amendment. He admits, despite everything, that he had to tweak -on the margin- the wording to stick to the French legislative framework.
On the other hand, another text written by ChatGPT is reproduced identically in the summary statement, which summarizes the essence of the amendment.
“It is important to guarantee adequate transparency in relation to the established algorithmic treatment, so that the interested parties can exercise their rights and request human intervention in case of reasonable doubt about the relevance or accuracy of the decision made. Therefore, this modification aims to strengthen the guarantees surrounding algorithmic processing in terms of public security, preserving the effectiveness of the latter in risk detection”, specifies ChatGPT, on the National Assembly site.
Despite this impressing initiative, Jean-Félix Acquaviva does not plan to use artificial intelligence again for his parliamentary work.
The signed ChatGPT amendment must be debated on March 21 or 22, before its possible adoption by the deputies.
Source: BFM TV
