AI for Advancing Biology
Throughout history, scientific progress has been defined by the tools we build to extend our perception. Today, artificial intelligence stands as our newest instrument -not merely as a computational engine, but as a revolutionary lens for deciphering the nature of reality. This talk will cover the philosophy and approach of Google DeepMind's AI for Science program. In particular, I will discuss our approach to selecting "root node" problems at different scales of biology - from predicting the shape and function of proteins to understanding the effect of genomic variants - and how we leverage AI toward their solution. I will conclude by reflecting on the transition toward open-ended, agentic AI systems that actively generate novel hypotheses, and their implications for the future of scientific inquiry.
Pushmeet Kohli, Ph.D.
Google Deepmind
Pushmeet Kohli is Vice President, AI for Science at Google DeepMind & Google Cloud’s Chief Scientist. He founded and leads the AI for Science and Strategic Initiatives Unit, which uses AI for breakthroughs in biology, materials,
fusion, math, software, and cybersecurity. His team developed AlphaEvolve for accelerating AI research and SynthID for safe AI deployment.
Key achievements include the Nobel Prize-winning AlphaFold model for protein structure prediction, one of the most cited AI biology papers (40,000+ citations). AlphaFold was the basis for the DeepMind spin-off, Isomorphic Labs. Kohli's team also pioneered large language model-based AI agents like AlphaEvolve and AI Co-Scientist, leading to discoveries in math and biology. Recently, Kohli’s team released AlphaGenome Atlas, which characterises every single-letter DNA variant possible in the human genome.