Google Launches Gemini for Science, Compressing Years of Research Into Days With Multi-Agent AI
Summary
Google launches Gemini for Science, a powerful multi-agent AI platform compressing years of scientific research into days, with tools already tackling liver disease, aging, and virus transmission alongside major enterprise and government partners.
Key Points
- Google is launching Gemini for Science, a suite of experimental AI tools built across Google Research, DeepMind, Google Cloud, and Google Labs that compresses scientific work from months or years into days.
- The centerpiece is Co-Scientist, a multi-agent AI system built on Gemini where specialized agents generate, critique, rank, and refine research hypotheses, with early applications spanning liver disease, aging research, and animal-to-human virus transmission studies.
- The platform includes four key components — Hypothesis Generation, Computational Discovery, Literature Insights, and Science Skills — and is already being deployed with enterprise partners like BASF and Bayer Crop Science, as well as US National Labs under the Department of Energy's Genesis Mission.