As fears about AI's power dominate the news, Yann LeCun, Meta's Chief AI Scientist, dismissed worries about it threatening humanity.
In a recent interview with The Wall Street Journal, LeCun called the idea of AI's impending world domination "complete B.S." He also said AI can't surpass human intelligence.
Should We Be Afraid of AI’s Growth?
LeCun's argument is deeply rooted in his unparalleled understanding of AI. He is one of the "Godfathers of AI," like Geoffery Hinton and Yoshua Bengio. So, his dismissal of AI's potential threat is not to be taken lightly. Many in tech warn about AI becoming super-intelligent. But, LeCun's expertise leads to a different conclusion.
The Limitations of Current AI Systems
Yann LeCun says current AI models, like LLMs, need general intelligence. He says AI models like ChatGPT and Grok are just language prediction models.
These excel at generating human-like text. But, they need better reasoning, planning, and understanding of the physical world. He emphasizes that these systems could be more intelligent but skilled at mimicking human language.
A Different Approach to AGI
LeCun says today's AI models will lead to something other than AGI (Artificial General Intelligence). However, he isn't against the idea of AGI and thinks it could be possible but via different methods.
As head of Meta's AI Group, LeCun is not just dismissing fears. He is working to develop an AI that can better understand the real world by analyzing video data. He believes this kind of research may lead to more advanced AI systems that eventually think like humans, offering a hopeful vision of AI's future.
Contrasting Views in the Tech Community
Meta's Chief AI Scientist’s views aren't just different; they are at the heart of a lively debate in the tech community. While others, like Elon Musk and Sam Altman, worry about AI's risks, LeCun says over-regulation will slow progress and innovation.