- Europe should keep AI models open source or risk falling behind, said Meta’s chief AI scientist.
- Open-source AI models allow for the free and open sharing of software to anyone for any purpose.
- Yann LeCun said that an open-source model allows everyone to benefit because progress is faster.
Europe should keep artificial intelligence models open source or risk falling behind, said Meta’s chief AI scientist.
Yann LeCun said that Europe and some countries are trying to make open-source models illegal because they want to stay ahead of political rivals, which is a “huge mistake.”
He made the comments during his presentation at the AI Action Summit in Paris on Monday.
“When you do research in secret, you fall behind,” the French-American computer scientist said. “The rest of the world will go open source and will overtake you. That’s currently what’s happening”
Open-source AI models allow for the free and open sharing of software to anyone for any purpose.
LeCun has been a strong advocate for open-source large language models and has reiterated that these systems should not be controlled by a small number of people or companies. He said that an open-source model allows everyone to benefit because progress is faster.
“We cannot afford to have those systems come from a handful of companies from the West Coast of the US or China,” LeCun said on Monday.
His comments follow the late-January release of an AI model from DeepSeek, an emerging Chinese AI startup, which rattled US tech and AI companies. Third-party tests showed the model outperformed its peers from OpenAI, Meta, and other top developers, and the company said it was built for less money.
DeepSeek’s R1 model is open source, which lets others download and build on top of it.
“DeepSeek has profited from open research and open source (e.g. PyTorch and Llama from Meta),” the chief scientist wrote in a Threads post in January. “They came up with new ideas and built them on top of other people’s work.”
Meta’s AI models, called Llama, are mostly open-source, which LeCun has advocated for at the company. OpenAI, originally founded as an open-source AI company, has more recently shifted to closed-source models.
European AI companies that use open-source models include French startup Mistral and Germany’s Aleph Alpha.
Both companies have criticized European proposals to regulate foundational model makers. Legislators in their home countries and in Italy have pushed for a framework in which model makers are allowed to self-regulate so they can compete with US tech giants.
European Union’s Artificial Intelligence Act, approved in 2024, aims to combat the risk associated with powerful AI technology. Most recently, there has been debate about how to regulate foundation models like large language models under the act.
