Amine Benyamina has finally called for wariness toward AI as a computing object, but also because of its concentration in the hands of a few large technology companies pursuing commercial aims.
The commission has brought together, since February, psychiatrists, legal experts, economists and scientists.
He cited the example of the May 2026 legislative elections in Scotland. According to the researcher, “35% of the information distributed by AI to people who were due to vote was erroneous.”.
Amine Benyamina, co-chair of the commission on generative artificial intelligence and vulnerabilities, called for “not running into the wall.” On France Inter, he set out the findings drawn from the work of this body. According to him, researchers have identified “vulnerabilities and risks” linked to the use of generative AI. These weaknesses, he indicated, can also weigh “against our model of society.”.
The warning thus concerns both the individual effects of the tool and the framework in which it is developed, as electoral deadlines approach in which speech produced by AI appears, according to him, less debated.