This is a translation of the original article in Portuguese. Read the original (Português).
France’s Senate is scheduled to vote on November 19 on a bill on artificial intelligence and copyright, authored by Senator Laure Darcos. The central point is the presumption that protected works were used by AI models, a rule that the French company Mistral AI rejects.
According to the French newspaper Les Echos, lobbyists from technology companies and representatives of the cultural sector are working on both sides to influence the text before the vote. The dispute pits those who develop the systems against those who create the content used in training.
In practice, a presumption of use reverses the burden of proof. Instead of authors and artists having to demonstrate that their works fed a model, it would fall to the AI company to contest that assumption. Mistral AI opposes this mechanism, according to the newspaper.
The issue matters to Brazil, where the debate over how to pay authors and artists when their works are used by AI systems is also open. The Brazilian discussion involves the same tension between encouraging the development of the technology and protecting those who produce music, text, images and other content.
A rule approved in France could serve as a reference for other countries seeking a balance in this area. For Brazilian creators and companies in content, creative work and technology, the outcome in the French Senate will be watched as a signal of which path may gain strength.
The text still must pass a vote by senators, and the result will determine whether the presumption of use is kept in the proposal.
Reporting from Les Echos (Lei sobre IA e direitos autorais divide setor de tecnologia e cultura, p. 25 (França)).
Deixe um comentário