- The chief artificial intelligence scientists in Meta said that Europe should keep the muddles of open or risk.
- The open source AI models allow the free and open participation of programs for anyone for any purpose.
- Yan said that the open source model allows everyone to benefit because progress is faster.
The head of Amnesty International in Mita said that Europe should keep the models of open -source artificial intelligence or risk.
Yan said that Europe and some countries are trying to make open source models illegal because they want to stay at the top of political competitors, which is a “big mistake.”
He made comments during his show at the AI Action summit in Paris on Monday.
“When you search secretly, you are behind the knees,” said the French -American computer world. “The rest of the world will do an open source and it will surpass you. This is what is happening now.”
The open source AI models allow the free and open participation of programs for anyone for any purpose.
Lecun has been a strong defender of large open source language models, and he has reiterated that these systems should not be controlled by a few people or companies. He said that an open source model allows everyone to benefit because progress is faster.
“We cannot afford the costs of the presence of these systems from a handful of companies from the western coast of the United States or China,” he said to be on Monday.
His comments follow the release of the late January of the artificial intelligence model from DibsicAnd it is a Chinese emerging company of artificial intelligence, which The United States shook technology and artificial intelligence companies. Third -party tests showed that the model outperformed its peers from Openai, Meta and other developers, and the company said it was built with less money.
Deepseek’s R1 model is open source, which allows others to download and build on.
“Dibsic has benefited from open research and open source (for example Pytorch and Llama from Meta),” Senior Scientists wrote in the Publishing Publishing in January. “They have reached new ideas and built them at the head of others.”
Meta models of artificial intelligence, called Llama, are mostly open source, which was called to be in the company. Openai, which was originally established as an open source AI, was recently transferred to closed source models.
European artificial intelligence companies that use open -source French emerging models include Mistral and Aleph Alpha in Germany.
Both companies have criticized European proposals for organizing constituent models. Located in their countries of origin and Italy pushed a framework that the self -regulation makers allowed to compete with American technology giants.
The European Union’s artificial intelligence law, which is accredited in 2024, aims to combat risk associated with strong artificial intelligence technology. Recently, there has been a debate on how to organize basic models such as large language models under law.
