One of the most prominent figures in artificial intelligence has a blunt message: today's AI is not really intelligent.
According to the BBC, leading AI researcher Yann LeCun argues that current systems fall short of genuine understanding. "We don't have robots that are nearly as good at understanding the physical world as a rat," he told the BBC — a striking comparison from someone regarded as one of the field's pioneers.
The point cuts against the hype around chatbots and generative tools. These systems can produce fluent text and images, but LeCun's argument is that fluency is not the same as grasping how the physical world actually works — the kind of everyday, common-sense reasoning that even simple animals manage effortlessly.
The BBC reports that LeCun is not just critiquing from the sidelines. He now has a start-up focused on developing a more flexible AI system — one aimed at moving beyond the limitations he describes toward machines that can better understand their surroundings.
Why does this matter? LeCun is not a fringe skeptic; he is, by the BBC's description, one of the leading figures in the world of AI. When someone at the center of the field says the current approach is "not smart," it reframes the debate about how close we really are to human-like machine intelligence — and suggests the next breakthroughs may require a different path than simply scaling up today's models.