Two AI-based science assistants succeed with drug-retargeting tasks
Summary
Two AI-powered science assistants are described in Nature papers, one Google’s Co-Scientist and another from FutureHouse, designed to help scientists generate and test hypotheses from large biological data sets. The systems act as agentic tools that can search literature, suggest mechanisms, and propose drug targets, while keeping humans in the loop for validation and safety considerations. The article emphasizes both the promise of accelerating hypothesis generation and the current limits and need for human oversight.