Finance Associated Press, October 14th (Editor Huang Junzhi) Meta's chief artificial intelligence scientist, known as the "AI godfather" Yann LeCun, does not seem concerned about artificial intelligence posing a threat to humans.

During a recent interview, he elaborated on his views, and when asked about whether AI is smart enough to pose a threat to humans, he said: "Please forgive my poor French, but this is complete nonsense."

Yann LeCun believes that current large language models lack some key abilities, such as long-term memory, reasoning, planning, and understanding the physical world. In his view, AI only proves that you can manipulate language, but is not intelligent and will never bring true general artificial intelligence (AGI).

In fact, this is not the first time Yann LeCun has expressed such views. He has previously publicly expressed his doubts, such as posting on X, "Before we worry about controlling super-intelligent AI, we need to design a smarter system first."

He believes that it is not possible to create a general artificial intelligence (AGI) comparable to human intelligence using existing advanced large language models, as these models can only accurately answer prompts when provided with the correct training data.

Overall, Yann LeCun believes that although current large language models excel in natural language processing, dialogue understanding, interactive conversations, and text creation, they are still just a "statistical modeling" technology, completing tasks by learning statistical patterns in data, fundamentally lacking true understanding and reasoning abilities.

However, this does not mean he completely doubts artificial intelligence. He just believes new methods are needed. For example, he pointed out that Meta's foundational AI research team is studying how to digest real-world videos.

In 2018, Yang Likun and two colleagues - Montreal Institute of Learning Algorithms Director Yoshua Bengio and former Google Vice President Geoffrey Hinton, jointly won the Nobel Prize in Computer Science - Turing Award for 'making deep neural networks a key component of computation'.

It is also because of this that the three computer scientists mentioned above are all honored as 'fathers of artificial intelligence'. Their groundbreaking work has paved the way for the artificial intelligence models we use today, such as OpenAI's ChatGPT and Anthropic's Claude.

But it is worth mentioning that the 'AI fathers' have different views on the above issues. Just last week, also known as 'AI father', Bengio warned that we are creating monsters (artificial intelligence) more powerful than humans.

"Wisdom gives power, and whoever controls this power - whether at human level or above human level, will become very very powerful. In general, technology will be used by those seeking more power, such as those seeking economic dominance, military dominance, political dominance. Therefore, we need to be very careful before creating technologies that could dangerously centralize power." he said.