Musk was criticized when he said 'AI is about to be smarter than humanity'

Elon Musk thinks AI could be smarter than all humans by 2029, but AI Director Meta Yann LeCun disagrees.

Last weekend, futurist Ray Kurzweil commented on the X platform that AI will be smarter than humanity by 2029. He said Stanford University organized a conference, inviting hundreds of people around the world to come. AI predictions. "Most think this will happen, but not in 2029, it will take about 100 years," he said.

Later, Musk re-shared the above message and commented: "AI may be smarter than any human next year. By 2029, AI may be smarter than all humans."

Meanwhile, Yann LeCun, Meta's chief AI engineer and one of the four leading professors who laid the foundation for the development of artificial intelligence, said that the scenario of AI being smarter than humans will not appear. as soon as Musk predicted. He used the self-driving car platform as an example.

"If that were true, we would have an AI system that would teach itself to drive a car with 20 hours of practice, something any 17-year-old teenager would do. In fact, we still do not have a fully automated driving system and reliable, even though Musk's system has had millions of hours of labeled training data," LeCun commented on X.

Musk did not respond. Previously, he repeatedly promised to bring fully automated driving features to Tesla cars, but so far no model has achieved that ability.

Yann LeCun (left) and Elon Musk. Photo: UCLA/Reuters

In the past, Musk has repeatedly said "AI will be dangerously strong", they can develop to a kind of super intelligence and overpower human systems. "The possibility of AI destroying humanity is very small, but not zero," he told the WSJ in May 2023.

Meanwhile, LeCun is one of the scientists who considers AI not a real threat. "The intelligence level of AI is far below human intelligence," he said at an event in Dubai last month, according to The Observer .

The French-American scientist assessed: "Why are systems not even as smart as a cat? Cats can remember, understand the physical world, can plan their complex actions . They do it better than the largest LLM language models today".

At the VivaTech conference in Paris last June, he also said AI dominating humans is an unrealistic worry. For example, he said chatbots like ChatGPT still have a major limitation: they are only trained on text and cannot fully capture human experience. Therefore, he believes that many experts only try to create fear around the threat of AI.