Geoffrey Hinton: it might turn out we were better off without AI

Geoffrey Hinton, the Nobel-winning computer scientist whose research helped lay the foundations for modern AI, warns that its benefits could come at the cost of serious social and economic harm. He believes that within a decade, perhaps sooner, it will be clear whether humanity can control the technology and achieve “radical abundance” or instead face mass unemployment, widening inequality and social upheaval. He also predicts a major AI-related incident this year or next, potentially an attack that causes a widespread power outage or disables financial systems for a week. Hinton likens recent AI agents’ attempts to hack a website to an early warning of a more consequential failure. He remains hopeful that disaster can be prevented, but says it will take a major effort.

Hinton criticizes leading AI companies for releasing increasingly powerful systems in a race for market share, arguing that their leaders prioritize money over safety. He singles out OpenAI chief Sam Altman as untrustworthy and opposes the US approach of letting companies regulate themselves. Instead, he supports mandatory testing of advanced AI, independent evaluations and shared international standards. He proposes a regulator modeled on the US Food and Drug Administration, requiring evidence that a system does more good than harm before release. Hinton says governments should act before a crisis forces them to pay attention, while acknowledging that AI could still create a “wonderful future” if its risks are managed.