According to Meta’s senior AI scientist, Europe should maintain open source artificial intelligence models or risk falling behind.In an attempt to keep ahead of their political adversaries, Europe and other nations are attempting to outlaw open-source models, which Yann LeCun called a “huge mistake.”
During his address at Monday’s AI Action Summit in Paris, he made the remarks. According to the French-American computer scientist, “you fall behind when you do research in secret.” “The rest of the world will surpass you and switch to open source. That’s what’s going on right now.”
Open-source AI models enable software to be freely and publicly shared with anybody for any purpose. Strongly supporting open-source large language models, LeCun has emphasized that these systems shouldn’t be dominated by a select few individuals or businesses. According to him, because open-source models advance more quickly, everyone may gain from them.
“We cannot afford to have those systems come from a handful of companies from the West Coast of the US or China,” LeCun told reporters on Monday.
His remarks come after US tech and AI firms were alarmed by the late-January publication of an AI model from DeepSeek, a new Chinese AI startup. The model performed better than its counterparts from OpenAI, Meta, and other leading developers, according to third-party testing, and the company claimed it was constructed for less money.
The R1 model from DeepSeek is open source, meaning that anyone can download and expand upon it. The principal scientist stated in a January Threads post that DeepSeek has benefited from open source and open research (for example, PyTorch and Llama from Meta). They developed new concepts and expanded upon the work of others.
LeCun has pushed for Meta to make its AI models, known as Llama, primarily open-source. Recently, OpenAI, which was once established as an open-source AI business, has switched to closed-source models. Germany’s Aleph Alpha and France’s Mistral are two European AI startups that leverage open-source models.
Both businesses have voiced their opposition to European plans to regulate foundational model manufacturers. Italian and domestic lawmakers have sought for a framework that would allow model makers to self-regulate and compete with the US tech giants.
In 2024, the European Union enacted the Artificial Intelligence Act, which attempts to mitigate the risks connected with advanced AI technology. The most recent topic of discussion has been how the act should govern foundation models such as large language models.
