Advances in artificial intelligence are rapidly transforming the scientific enterprise, offering powerful new methods for generating hypotheses, designing experiments, and interpreting data. I will discuss the promise of AI to accelerate scientific discovery, sharing examples that illustrate emerging patterns of innovation driven by the rapid scaling of cycles of generation and testing.

I will draw from case studies in materials science and biology to highlight these trends. Finally, I will address growing concerns around dual-use risks, where the same technologies that drive beneficial advances can also lower barriers to the creation of harm.