Meta's Yann LeCun Says He Doesn't Believe that AI Can Threaten Humanity

Yann LeCun, who is a senior researcher at Meta, says that he doesn't think that AI can become intelligent in the future. Previously, LeCun emphasized that we don't need to worry about AI reaching superintelligence just yet because, before that happens, we still need to figure out how to design AI that surpasses even simpler forms of intelligence—like that of a house cat. His point is that we're far from creating AI systems that are even as "clever" as animals, let alone superintelligent beings.

He gave an interview to the Wall Street Journal and when asked about what he thinks about AI becoming intelligent and powerful enough to become a threat to humanity, he replied that it is complete nonsense. LeCun added that right now a lot of LLMs do not have everything that they need to become a threat to humanity. They do not have the capabilities of reasoning, understanding, planning and knowledge about the real world. We can simply manipulate the LLMs but this can never become AI super intelligence.

This doesn't mean that he neglects AGI completely. He just said that new ideas and approaches are needed and he suggested to study his work with Meta's AI research team to make sense of his thoughts about this whole topic.

Image: DIW-Aigen

Read next: Apple Simplifies Email Management with New Primary Email Option in iOS 18.1

Previous Post Next Post