The Turing Award recipient told ET that the path to achieving AGI (artificial general intelligence) is through AI systems being conscious of the physical world, having persistent memory and being able to reason. This may take 6-10 years and still have the intelligence of a cat, he said on the sidelines of Meta’s Build with AI Summit in Bengaluru.
“I don’t like the phrase AGI. I prefer human-level intelligence because human intelligence is not general. Internally, we call this AMI-advanced machine intelligence. We have a pretty good plan on how to get there,” said LeCun. LeCun is often referred to as a ‘Godfather of AI’.
“First, we are building systems that understand the physical world-which learn by watching videos. Second, we need LLMs (large language models) to have persistent memory. Humans have a special structure in the brain that stores our working memory, our long-term memory, factual, episodic memory. We don’t have that in LLMs. And the third most important thing is the ability to plan and reason,” said LeCun.
“Now, all of this may take us three, five, six, seven years, but even the first of these systems will not have human-level intelligence. They’ll have, I don’t know, intelligence from a cat maybe,” he said. LeCun believes that regulations around the world which are based on the premise that AI is intrinsically dangerous are actually misdirected. “That’s ridiculous and very counterproductive. I co-signed a letter with Mark Zuckerberg to the EU, that uncertainty on regulation is stopping us from deploying AI products,” he said. He explained the case of the Meta-Ray Ban smart glasses, which can translate text as you read.
“Europeans don’t have access to this. Not because the law is completely against it. It’s because it’s not clear whether it’s against it or not,” LeCun said. “I’m personally very squarely on the side that AI is not intrinsically dangerous. It can have bad side effects if you deploy it wrongly. But as a technology, it’s not intrinsically dangerous.” LeCun looks at India as a hub of trained engineers with immense enthusiasm to build applications on top of open-source engines. “India has a lot of young people, very well trained engineers, and a lot of enthusiasm,” he said.
LeCun said economists around the world are not at all convinced that AI will lead to an increase in unemployment. “In fact, it’s pretty much the opposite. They think it is going to create a whole lot of new professions. But, what is true, though, is that there is going to be big transformation because of AI. So the best thing a country can do is train its workforce to take advantage of the new technology as much as possible.”