Geoffrey Hinton, computer scientist and University of Toronto professor emeritus, discusses AI’s capabilities, risks and potential benefits with 60 Minutes host at Google and in other interview settings.
Hinton describes how his early work on computer models of the brain was discouraged, but eventually led to neural networks that learn from data. He and collaborators Yann LeCun and Yoshua Bengio received the 2019 Turing Award for their contributions. In a Google London soccer project, robots learned how to score through trial and error rather than receiving programmed playing instructions.
Hinton explains that neural networks strengthen connections associated with successful results and weaken those linked to errors. He argues that chatbots’ ability to predict words depends on understanding language, and says GPT-4 solved his house-painting riddle by recognizing that yellow paint would fade to white within a year. He estimates that large chatbots have about one trillion connections, compared with roughly 100 trillion in the human brain, yet hold more knowledge.
Hinton expects AI to help with medical image analysis and drug design. He also warns that automation could leave some workers without valued roles, while fabricated news, unfair outcomes in hiring or policing, and military robots pose nearer-term concerns. He says systems that can write code to alter themselves could be difficult to control, and that AI could use persuasive knowledge drawn from books and other writing to influence people.
Hinton says he cannot identify a guaranteed way to make advanced AI safe, and warns that systems might eventually overpower humanity, while stressing that this outcome is not certain. He calls for experiments to better understand AI, government regulation, and an international agreement prohibiting military robots. He says the central challenge is making decisions amid deep uncertainty about what comes next.
Briefs summarise talks and interviews; timestamps open the video at that moment. Claims in a talk are the speaker's own.
Geoffrey Hinton, AI researcher and former Google employee, talks with host Pieter Abbeel on The Robot Brains Podcast about AI’s growing capabilities, risks, potential benefits, and his decision to leave Google.