Gemini for Science: AI experiments and tools for a new era of discovery

2026-05-19 · Source: Google DeepMind · Field: Science & Research — Artificial Intelligence & Machine Learning, Research Methodology & Innovation, Life Sciences & Biology · Depth: Advanced, long

Summary

Google has introduced "Gemini for Science," a new suite of AI tools and experiments designed to accelerate and enhance scientific discovery. This initiative includes three experimental prototypes available on Google Labs: Hypothesis Generation, built with Co-Scientist, which simulates the scientific method to generate and evaluate hypotheses with verified claims; Computational Discovery, powered by AlphaEvolve and ERA, an agentic research engine that tests thousands of code variations for complex modeling in fields like solar forecasting; and Literature Insights, utilizing Google NotebookLM, which structures scientific literature for analysis and artifact creation. Additionally, "Science Skills" integrates over 30 life science databases, including UniProt and AlphaFold Database, into platforms like Google Antigravity, significantly reducing time for structural bioinformatics and genomic analyses. These tools are being gradually rolled out, with enterprise versions already in private preview with partners like BASF and Klarna, and research validated by papers in Nature and collaborations with over 100 institutions.

Key takeaway

For research scientists and R&D leads aiming to accelerate discovery, Google's Gemini for Science provides agentic AI tools that can significantly streamline your workflows. You can utilize Hypothesis Generation to rapidly ideate, Computational Discovery for parallel experimentation, and Literature Insights for efficient synthesis of scientific papers. Explore registering for access via labs.google/science or Google Cloud to integrate these capabilities and multiply your team's research output.

Key insights

Gemini for Science leverages AI agents to automate complex research tasks, significantly accelerating scientific discovery and exploration.

Principles

• General AI agents empower broad scientific research.
• AI multiplies human ingenuity by automating tasks.
• Scientific rigor demands verifiable, cited claims.

Method

AI agents define research challenges, generate and debate hypotheses, verify claims, and test thousands of code variations in parallel. They also synthesize literature into structured tables for analysis and artifact creation.

In practice

• Utilize AlphaEvolve for supply chain optimization.
• Apply Co-Scientist to accelerate research ideation.
• Employ Science Skills for rapid genomic analysis.

Topics

• Gemini for Science
• AI Agents
• Scientific Research Automation
• Hypothesis Generation
• Computational Discovery
• Structural Bioinformatics
• Genomic Analysis

Best for: AI Scientist, Research Scientist, Director of AI/ML

Related on AIssential

• Gemini for Science: AI experiments and tools for a new era of discovery — Gemini for Science uses general AI agents and specialized tools to accelerate scientific discovery by automating complex research tasks.
• Google launches Gemini for Science at I/O 2026 — Google's Gemini for Science suite uses AI to accelerate scientific discovery from hypothesis generation to literature review.
• Gemini for Science is here. 🧬 — Gemini for Science uses AI to streamline scientific research tasks and accelerate discovery.
• Google I/O showed how the path for AI-driven science is shifting — Google is shifting its AI for science focus from specialized tools to agentic, LLM-based systems for autonomous research.
• Co-Scientist: A multi-agent AI partner to accelerate research — Co-Scientist is a multi-agent AI system that iteratively generates, debates, and refines scientific hypotheses to accelerate discovery.
• Google unveils ERA AI to accelerate expert-level scientific research — Google's ERA AI, powered by Gemini, accelerates scientific discovery through automated code generation and experimental refinement.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Google DeepMind.