Google Deepmind's AI Co-Scientist now plans experiments, runs lab equipment, and writes scientific papers
Not a chatbot. Google Deepmind's AI Co-Scientist is now planning experiments, operating lab hardware, and writing papers—autonomous end-to-end research across three disciplines.
Why it matters
A multi-agent system deployed doing real, experimentally validated scientific work marks a major shift from lab assistant (hypothesis generator) to autonomous researcher. This is agents as a category proving capability in high-stakes, measurable domains—practitioners need to track how this scales.
The key facts
6 to knowCo-Scientist expanded from hypothesis generator to full research workflow (planning, execution, writing)
Multi-agent system built on Gemini architecture
Deployed across three disciplines: materials synthesis, medical AI architecture development, and one unspecified domain
Delivered experimentally validated results (not simulated or theoretical)
Integrated directly into physical lab (operates equipment autonomously)
Published Aug 28, 2026
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: Google Deepmind has expanded Co-Scientist from a hypothesis generator into a research system that's integrated into the lab. Across three disciplines, from materials synthesis to the autonomous development of a medical AI architecture, the Gemini-based multi-agent system delivered experimentally…