Recorded 10 February 2026. Richard Sutton of the University of Alberta presents "The future of AI" at IPAM's AI for Science Kickoff. Abstract: In this talk, I will make three big points. First, although AI today is based on learning from human data, the AI of the future will be based on learning from experience, which will be much more powerful because it can continually learn new knowledge. Second, the politics of AI parallels the politics of people; in both cases we should avoid centralized control based on fear/safety and seek decentralized cooperation. Third, AI can be viewed as the inevitable next big step in the development of the universe, in which case it should be embraced with courage, pride, and a sense of adventure. Learn more online at: https://www.ipam.ucla.edu/programs/special-events-and-conferences/ai-for-science-kickoff-2026/?tab=schedule About AI for Science Kickoff 2026: This inaugural event brings together the pioneers who are defining how AI will accelerate scientific discovery — from Nobel and Fields Medal laureates to the leaders shaping AI innovation across academia, research labs, and industry. The event features keynote talks by leading AI Scientists and Mathematicians, as well as panel discussions focusing on perspectives on AI from three sides: Mathematics, Higher Education, and Industry. This event is organized jointly by IPAM, the UCLA Division of Physical Sciences, the SAIR Foundation and the World Leading Scientists Institute.