Yann LeCun positioned himself as a strong advocate for sovereignty at VivaTech. The French researcher and leading figure in artificial intelligence (AI) defended governments’ efforts to secure technological sovereignty in the field, just days after a U.S. decision suspended foreign users’ access to certain tools developed by Anthropic.
“I have spoken with several governments around the world. They all want to ensure their sovereignty in AI, and I believe they are right,” he said on the main stage of VivaTech in Paris, addressing an audience of technology entrepreneurs.
“This is extremely important because, before long, virtually all of our information consumption will be mediated through AI assistants,” continued the former Chief AI Scientist at Meta, the parent company of Facebook and Instagram.
According to LeCun, having access to “a wide diversity of AI assistants” and “multiple sources of information” is just as important as maintaining media pluralism.
“The only way to achieve this would be through an open and freely available foundation model on top of which anyone could build their own specialized assistant, tailored to their language or languages, culture, value system, political views, and interests,” he emphasized, referring to the Tapestry project.
The initiative, which he launched in April together with the AI Alliance coalition, aims to develop an open and sovereign AI ecosystem.
“Mitigating the Risks of AI”
The issues of AI security and sovereignty have become central concerns in Europe after the Trump administration ordered U.S. AI startup Anthropic last week to suspend access to its two most powerful models, Claude Fable 5 and Mythos 5, for “all foreign nationals,” citing national security concerns.
The French scientist, who also founded the startup AMI Labs, criticized Anthropic’s decision to make its Mythos model—used to identify cybersecurity vulnerabilities—available only to a select group of companies for safety reasons.
“There is a great deal of arrogance and a superiority complex in the idea that only a few people are capable of controlling AI, while the uninformed masses should not have access to it,” he argued.
According to LeCun, the use of AI tools “does involve risks, and those risks must be mitigated,” but restricting access to a select few “is nothing more than a form of control.”
