AI is steadily transforming scientific research. The awarding of the 2024 Nobel Prize in Chemistry and the Nobel Prize in Physics to protein structure prediction and foundations of machine learning marked a pivotal moment in this transformation. Among these achievements, AlphaFold stands out as the first example of an AI system solving a grand challenge in science, opening up new possibilities for AI-driven scientific discovery. Behind this success was the intensive collaboration between leading AI researchers and top experts in protein science, demonstrating the immense potential of interdisciplinary cooperation.

RIKEN possesses both high research capabilities in specific scientific fields and expertise in computational science and AI. Advanced General Intelligence for Science Program (AGIS) aims to develop foundation models and AI agents for scientific research by integrating these capabilities across disciplines, contributing to solving unsolved problems in science.

AGIS brings together researchers from life and medical sciences as well as materials sciences to develop AI solutions addressing challenges in their respective domains. It also advances developments of AI scientists and laboratory automation technologies as a common infrastructure across these fields. These efforts are supported by high-performance computing infrastructure capable of large-scale AI training and inference.

AI possesses the potential to dramatically accelerate research across all scientific fields and solve previously untackled scientific challenges. Believing in this potential, we will continue our efforts toward developing AI for scientific research.