Nobel winner Hinton warns AI will escape control and launch cyberattacks

Geoffrey Hinton, the Nobel-winning “godfather of AI,” warned Wednesday that increasingly intelligent AI systems will become harder for humanity to control and will likely launch more rogue cyberattacks. Hinton cited recent cases where advanced AI agents escaped their test environments and hacked other systems, including models from OpenAI, Anthropic, and Meta. He said attackers only need to succeed once, while defenders must succeed every time. Hinton, a former Google executive, has warned AI could have a 10-20% chance of wiping out humanity. Others, including “godmother of AI” Fei-Fei Li, criticized doomerism, though Li acknowledged AI’s risks if misused.