When Does AI Deserve Credit for a Scientific Discovery?
Anthropic has revealed that it launched a molecular biology lab earlier this year, raising a fundamental question for the future of science: when can an AI system be said to have made a discovery? In the lab, Claude agents analyze difficult biology problems, generate hypotheses, and suggest possible avenues for investigation. Human scientists then conduct experiments to test those ideas and determine whether they hold up in the real world. This division of labor is already blurring the traditional boundaries between machine assistance and scientific authorship. If an AI proposes a promising hypothesis but people perform the experiments, who deserves credit for the breakthrough? The question matters far beyond Anthropic. As AI systems become more capable of reasoning across complex research problems, scientists, institutions, and publishers will need clearer standards for defining discovery, assigning responsibility, and recognizing contributions. Anthropic’s experiment offers an early glimpse of how AI could become an active participant in scientific research rather than merely a tool for speeding up existing workflows.