Skip to content

AI co-scientist

A multi-agent system that drafts research hypotheses and plans

Visit official site

What AI co-scientist is

The system is described as a multi-agent AI system built on Gemini 2.0 that simulates reasoning in the scientific method. It generates new research hypotheses and experimental plans and is described as being in a research stage.

What you can do with it

  • Generate new research hypotheses and experimental plans
  • Use generation, reflection, ranking and evolution mechanisms to refine hypotheses
  • Interact in natural language to give feedback or seed ideas
  • Draw on Gemini 2.0 to integrate knowledge across disciplines

Getting started

  • Describe a research goal in natural language
  • Enter the goal into the system so it can parse it and produce a research plan
  • Let multi-agent collaboration generate several hypotheses and plans
  • Evaluate the hypotheses and give feedback for further refinement

Reference cases

  • Drug candidates proposed in AML drug repurposing research
  • Antifibrotic epigenetic targets proposed in liver fibrosis research
  • A gene transfer mechanism proposed in antimicrobial resistance research

Pros & cons

✓ What we like

  • Hypotheses are refined through multi-agent collaboration
  • Scientists can supply feedback or seed ideas in natural language

! What to watch out for

  • The material describes the system as being in a research stage
  • The described validation work centres on the biomedical domain

FAQ

What is the system built on?

The material states it is a multi-agent AI system built on Gemini 2.0.

How does a scientist work with it?

The listed steps are describing a research goal, receiving generated hypotheses, then evaluating and giving feedback.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review