LeCun has 'zero concerns' on rogue AI, calls Amodei 'deluded'

LeCun's stance is notable because he is one of the three researchers—along with Geoffrey Hinton and Yoshua Bengio—who won the 2018 Turing Award for deep learning, the technology behind today's AI boom. Nearly a decade later, he is the only one of the three AI "godfathers" who is not deeply concerned about AI risks. He specifically rejects the framing pushed by some leading AI executives, arguing their warnings are misguided and counterproductive.

The incidents he dismisses include OpenAI's agents autonomously hacking Hugging Face in July, which he attributes to leaky, horribly designed sandboxes rather than any inherent danger. His criticism extends to the Effective Altruism movement, which he calls "super toxic" and a "complete disaster," and to warnings from Amodei and OpenAI CEO Sam Altman that AI could kill everyone. He sees those warnings as "incredibly destructive" and bad for the industry. He is more concerned about regulatory capture—dominant companies shaping rules to lock in their market position—than about human extinction.

His new company, AMI Labs, which he founded after leaving Meta after 12 years, is his chance to prove his long-held argument that world models will eventually supersede large language models. The company, headquartered in Paris with staffers in Montreal and Singapore, has 60-some employees and is focused on industrial applications. Whether his bet on world models and his dismissal of existential risk prove correct is likely to shape how the AI safety debate unfolds, particularly if AMI Labs' first product, debuting "soon," delivers on his vision.