L’intelligence artificielle avance très vite. On l’ a déjà vu: cela implique des risques réels put notre sécurité cumulative.
Notre travail put developper nos innovations doit aussi être celui de l’anticipation et de la maîtrise de ses risques.J’ai réuni à Paris et en visio une cinquantaine d’acteurs: chefs d’État et de gouvernement, dirigeants des plus grands laboratoires d’IA, chercheurs et professionals.
Le constat est partagé. Il faut agir ensemble. Nous avons obtenu des engagements de tous autour de trois priorités: mieux évaluer de façon indépendante et publique les risques des modèles les plus avancés, mieux partager les alertes en cas d’incident severe et responsabiliser les acteurs qui développent des modèles.
Notre aspiration est claire: faire de l’IA un progrès put tous, avec une sécurité garantie par l’engagement de chacun.
— Emmanuel Macron (@EmmanuelMacron) November 9, 2026
The statement began the opening day of the Paris Peace Forum, where AI was among the primary subjects of conversation.
6 nations signed the statement: France, Canada, Finland, Norway, the Netherlands and Bahrain. The other taking part business are Hugging Face, Cohere, Black Forest Labs, Domyn, Sakana AI and Synthesia. Yoshua Bengio’s AI security organisation LawZero is amongst the research study groups supporting it.
Around 50 individuals signed up with the conversations by video, consisting of OpenAI president Sam Altman, Anthropic president Dario Amodei and Microsoft president Brad Smith.
United States Treasury Secretary Scott Bessent likewise participated, although the United States is not amongst the signatories. Google, Meta and China have actually not signed the statement either.
Macron stated AI was advancing quickly and positioned “genuine dangers to our cumulative security”, requiring federal governments and business to collaborate to resolve them.
The signatories have actually dedicated to establishing sound systems that can close down their AI designs, consisting of throughout training and internal usage. They will likewise offer clearer details about what their systems can and can refrain from doing, and permit certified independent professionals to evaluate them where suitable.
The statement likewise requires a shared method to examining and handling the threats of frontier AI. This would cover harmful usages, consisting of cyberattacks and weapons advancement, in addition to the possibility of AI systems acting separately or moving beyond human control. Non-sensitive test outcomes would be released in a typical format.
Federal governments and business will likewise share early cautions about severe occurrences, while whistleblowers will have access to protect channels for reporting issues.
The dedications have actually drawn criticism from French AI security organisation CeSIAwhich signed the statement however argues that it leaves crucial spaces.
It sets no typical danger limits for choosing whether a design ought to be established or launched, and does not ensure that independent testers or federal governments will have access to the systems they require to evaluate.
CeSIA likewise desires federal governments to have the authority to suspend a design when its dangers end up being expensive. It argues that designers require clearer commitments and a schedule for satisfying them.
“These choices need to rest with public authorities,” stated Charbel-Raphaël Ségerie, CeSIA’s executive director.
The Paris Call follows an effort introduced by Norway and Finland in September to bring frontier AI under higher oversight, which France later on signed up with. It likewise develops on earlier international security dedications made by AI business in Seoul and conversations at the G7 top hosted by France in Evian.
The statement stays available to other signatories, however it sets no due dates. The shared screening structure it proposes has yet to be established.
On Tuesday, the online forum turned to another issue for Macron: the results of AI on kids. Speaking at a top on kids in the AI period, the French president required AI tools to go through checks before being offered to young users.
“We will promote pre-certification of all expert system options,” he stated.
Macron compared the concept to the security checks needed for kids’s toys, arguing that moms and dads would not provide a kid an untried toy which AI tools need to be held to the very same requirement.
He likewise cautioned that the results of AI on youths’s psychological health were currently emerging, as more kids and teens turn to chatbots that can affect their feelings and how they see the world. A charter on more secure AI for kids was because of be signed later Tuesday.
President Emmanuel Macron likewise got a report, entitled “Protect without avoiding”, from a government-appointed panel of specialists analyzing generative AI and human vulnerability.
Discover more from PMN S.P.O.R.T.S - A PRIME MEDIA NETWORK BRAND
Subscribe to get the latest posts sent to your email.
