Yann LeCun at VivaTech: A Sovereign AI for Every State

Yann LeCun Defends Digital Sovereignty at VivaTech

During his participation at the VivaTech conference in Paris, Yann LeCun, a prominent figure in artificial intelligence, took a stand in favor of the digital sovereignty of states. The researcher emphasized the importance for governments to maintain sovereign control over AI technologies, especially after the United States recently restricted access to certain tools from Anthropic for foreign users. This decision has highlighted the stakes of technological sovereignty.

Addressing an audience of entrepreneurs and technology professionals, LeCun stated, “I have had discussions with several governments around the world. They all want to ensure their sovereignty in AI, and I think they are right.” He underscored that this sovereignty is essential, as the future of information consumption will be largely influenced by AI assistants. As a former Chief Scientist of AI at Meta, LeCun is well aware of the implications of this evolution.

The Importance of Diversity in AI Assistants

Yann LeCun also highlighted the necessity for a wide diversity of AI assistants to ensure a plurality of information sources, similar to that of traditional media. He explained that an open and free base model would allow anyone to develop their own AI assistant, tailored to their linguistic, cultural, and political needs. This vision is embodied in the Tapestry project, which he launched with the AI Alliance in April. This project aims to create an open and sovereign AI, enabling each nation to customize its technological tools according to its own values.

Security and Sovereignty Challenges

The issue of securing and ensuring the sovereignty of AI has become a priority in Europe, especially after the Trump administration ordered Anthropic to restrict access to its Claude Fable 5 and Mythos 5 models for national security reasons. This measure has sparked debates about the balance between security and access to technology.

Yann LeCun, who founded the startup AMI Labs, criticized Anthropic's decision to limit the use of its Mythos model, designed to detect cybersecurity vulnerabilities, to a select group of companies. He condemned what he sees as a form of arrogance and a superiority complex, where only a few privileged individuals would have the right to control AI. For LeCun, while the use of AI tools carries risks, it is crucial to mitigate them without excessively restricting access.