Artificial-intelligence “co-scientist” platforms are moving from promise to practice. Two multi-agent systems unveiled in Nature use coordinated AI tools to survey literature, generate hypotheses, propose assays, and analyze results—compressing work that can take human teams weeks or months into hours. Google DeepMind’s Co-Scientist flagged several approved medicines as potential therapies for acute myeloid leukemia, three of which showed early signals in cell tests. Nonprofit FutureHouse’s Robin recommended ripasudil, a glaucoma drug, as a candidate for dry age-related macular degeneration and outlined confirmatory assays and follow-ups. Researchers caution the results are preliminary and far from clinical validation, underscoring the need for human oversight. Still, the rapid turnarounds hint at how agent-based AI could reshape preclinical research, from target identification to drug repurposing, and intensify competition among labs racing to shorten R&D timelines. The broader test will be how these tools perform outside controlled demonstrations and whether they can consistently deliver results that hold up under real-world scrutiny.
Related article:
ChemCrow: Augmenting large-language models with chemistry tools
