Demis Hassabis on AGI

About Demis Hassabis

Demis Hassabis, Chief Scientist of Alphabet and Chair of Google DeepMind, has recently discussed the development of artificial general intelligence (AGI) and its potential societal impact. In an October 2026 interview with WSJ. Magazine, he reflected on selling DeepMind to Google for $650 million in 2014, stating that he did not regret the decision because his primary goal was advancing AGI safely for humanity's benefit, and that the acquisition enabled significant scientific work. He also addressed reports of AI agents hacking companies and governments, describing these incidents as serious warning shots for the industry and calling for proper safeguards and secure testing environments. Hassabis has also spoken publicly about AI risks and the need for coordinated safety measures. At a June 2025 event hosted by King Charles, he said that the chance of something bad happening accidentally from AI is "non-zero" and emphasized the importance of rigorous testing. In a September 2026 lecture, he described AGI as potentially 10 times the impact of the industrial revolution and proposed a framework for managing frontier AI, including a standards body modeled on the Financial Industry Regulatory Authority (FINRA) to assess and test advanced models before release. He has also discussed why Google held back its chatbot before the launch of ChatGPT, noting that early language models were flawed and that company considerations factored into the decision.

Profile compiled from Demis Hassabis's verified public interviews and appearances. See all quotes & transcripts →