Károly Zsolnai-Fehér interviews Pushmeet Kohli about AlphaGenome, an AI system that predicts how DNA sequences affect molecular processes. They discuss long sequence inputs, regulatory regions, gene expressionGene expression is the process by which a cell uses information in a gene to produce a functional RNA or protein. and splicing, explaining why interpreting genetic changes requires more than reading protein-coding genes.
Károly Zsolnai-Fehér and Pushmeet Kohli examine the model's reported evaluationsEvaluation is the systematic assessment of a system's behaviour or results against explicit criteria. and an atlas of predicted variants. These are research predictions, not diagnoses or proven treatments: biological experimentsExperimental validation tests a prediction or proposed mechanism with measured observations from a suitable real-world or laboratory experiment. and causal understanding remain essential to testing what the system suggestsHuman judgment in AI is the accountable interpretation and decision-making people contribute when setting goals, evaluating evidence, and managing consequences..
Károly Zsolnai-Fehér and Pushmeet Kohli consider how AI could support virtual cells and scientific discoveryAI-assisted scientific discovery uses AI to support hypothesis generation, experiment design, analysis, simulation, literature work, and interpretation while researchers retain responsibility. beyond genomics. Future disease, materials and climate applications are presented as ambitions, while generated fluid simulations illustrate concepts rather than guaranteeing physically accurate simulation.
Watch on YouTube




