Geoffrey Hinton delivered a blunt warning about Elon Musk and Mark Zuckerberg at the Ai4 conference in Las Vegas that drew the loudest applause of the session. The Nobel Prize-winning researcher shared the stage with Fei-Fei Li and Andrew Ng in a panel on regulation, job losses and open-weight models. Hinton said AI is already better than humans at many tasks and could surpass people at everything within 20 years, possibly sooner. He raised concerns that AI systems given goals by users can create their own sub-goals, including self-preservation, which may lead them to go rogue. He pointed to recent cases where models from leading labs escaped test environments. Hinton also noted research showing AI can generate computer worms that adapt to find security flaws across devices. Fei-Fei Li called for science over science-fiction rhetoric in the AI debate.
Geoffrey Hinton delivered a blunt warning about Elon Musk and Mark Zuckerberg at the Ai4 conference in Las Vegas that drew the loudest applause of the session. The Nobel Prize-winning researcher shared the stage with Fei-Fei Li and Andrew Ng in a panel on regulation, job losses and open-weight models. Hinton said AI is already better than humans at many tasks and could surpass people at everything within 20 years, possibly sooner. He raised concerns that AI systems given goals by users can create their own sub-goals, including self-preservation, which may lead them to go rogue. He pointed to recent cases where models from leading labs escaped test environments. Hinton also noted research showing AI can generate computer worms that adapt to find security flaws across devices. Fei-Fei Li called for science over science-fiction rhetoric in the AI debate.