Barry Barish: Accelerating Basic Science with AI. Advancing AI through Fundamental Research.
In these closing remarks, Barry Barish returns to a question at the center of scientific discovery: how do we know when we have truly discovered something? Looking back to the birth of modern science, he traces the development of the scientific method through Copernicus, Galileo, Kepler, and Newton, then explains how the 20th century added statistical confidence as a central tool for evaluating evidence. Barish argues that the AI era creates a new version of this problem. AI may improve simulations, experiments, synthesis, analysis, and discovery workflows, but the scientific community still needs a convincing way to establish confidence in AI-assisted discoveries. Just as five-sigma became a shared standard for many areas of science, he suggests that AI for science now needs an equivalent framework for trust, validation, and statistical meaning. Outline of this video: 00:00 Returning to the question of discovery 00:28 Why discovery has been Barry Barish's lifelong focus 01:08 How do we believe a discovery in the era of AI? 01:33 The historical roots of modern science 01:53 Copernicus, Galileo, Kepler, and Newton 02:31 Newton and the scientific method 03:01 Updating the scientific method for machine learning and AI 03:41 How statistics changed discovery in the 20th century 04:21 Five-sigma and scientific confidence 04:41 Entering the AI era 04:50 The challenge of knowing when AI has helped make a discovery 05:03 Why AI-assisted discovery still faces a validation bottleneck 05:32 AI scores and why they are not yet enough 05:58 Why this remains an unsolved problem for AI and science 06:08 The need for an equivalent of statistical analysis 06:36 A challenge for the field

This is not the AI we were promised | The Royal Society

Terence Tao: What's Next for SAIR Competitions

Yann LeCun: World Models: Enabling the next AI revolution

The Uncomfortable Truth About AI “Reasoning” | World Science Festival

A 28-year-old Steve Jobs gives a talk at the 1983 International Design Conference in Aspen

Ex-Google Insider: You're Not Ready For The Next Phase of AI

Is AI Hiding Its Full Power? With Geoffrey Hinton

Keynote: After the AI Hype – What’s Real, and What’s Next - Richard Campbell - 2026

The Future of Science With AI | Nobel Prize Dialogue London 2026

The Future of Education How AI is Transforming Critical Thinking

AI, Machine Learning, Deep Learning and Generative AI Explained

How To Think SO Clearly People Assume You're Brilliant

AlphaFold - The Most Useful Thing AI Has Ever Done

Can space and time emerge from simple rules? Stephen Wolfram thinks so. | World Science Festival

AI has hacked the code of human civilization | Yuval Noah Harari

The Hardest Questions in Physics | World Science Festival

How AI Cracked the Protein Folding Code and Won a Nobel Prize

What rebuilding AlphaGo teaches us about self-play, RL, and future of LLMs - Eric Jang

The Rise and Reckoning of AI | 2026 Isaac Asimov Memorial Debate

