Accelerating scientific discovery with Co-Scientist

Ускорение научных открытий с помощью Co-Scientist
Pushmeet Kohli, Demis Hassabis, Keran Rong, Nenad Tomašev, Gary Peltz, Tao Tu, Demis Hassabis, Ottavia Bertolli, Eeshit Dhaval Vaishnav, Juraj Gottweis, Yossi Matias, Alan Karthikesalingam, Vivek Natarajan, Katherine Chou, Khaled Saab, V. S. Dhillon, Guan Yuan, José R. Penadés, Alexander Daryin, Wei‐Hung Weng, Anatoly Myaskovsky, Annalisa Pawlosky, Tiago R. D. Costa, Vikram Dhillon, Tom Sheffer, Anil Palepu, Petar Sirkovic, Felix Weissenberger, Dan Popovici, Fan Zhang, Avinatan Hassidim, Amin Vahdat, K. Kulkarni, B. Lee, Yunhan Xu, Jan Freyberg, Juraj Gottweis, Wei-Hung Weng, Tao Tu, Petar Sirkovic, Grzegorz Głowaty, Felix Weissenberger, Alessio Orlandi, Dan Popovici, Anil Palepu, Ryutaro Tanno, Khaled Saab, Fan Zhang, Jacob Blum, Andrew Carroll, Kavita Kulkarni, Nenad Tomašev, Dina Zverinski, Ivor Rendulić, Elahe Vedadi, Florian Hasler, Luka Rimanić, Marina Boia, Ivan Budiselić, Ben Feinstein, Mathias Bellaiche, Jeremy Ratcliff, Katherine Chou, Avinatan Hassidim, Burak Göktürk, Amin Vahdat, Yuan Guan, Byron Lee, Tiago R. D. Costa, José R. Penadés, Gary Peltz, James Manyika, Yunhan Xu, Alan Karthikesalingam, Vivek Natarajan, Grzegorz Glowaty, Alessio Orlandi, Keran Rong, Ryutaro Tanno, Jacob Blum, Andrew Carroll, Dina Zverinski, Ivor Rendulic, Elahe Vedadi, Florian Hasler, Luka Rimanic, Marina Boia, Ivan Budiselic, Ben Feinstein, Mathias Bellaiche, Tom Sheffer, Jan Freyberg, Jeremy Ratcliff, Ottavia Bertolli, B. Gokturk, Eeshit Dhaval Vaishnav, James Manyika
2026-05-19

antimicrobial resistancedrug repurposingmulti-agent AI systemscientific hypothesis generationtest-time compute scaling
Abstract Scientific discovery is driven by scientists generating hypotheses for complex problems that undergo rigorous experimental validation. To augment this process, we introduce Co-Scientist, a multi-agent artificial intelligence (AI) system built on Gemini for structured scientific thinking and hypothesis generation. Co-Scientist aims to help scientists discover new original knowledge. Conditioned on their research objectives and previous scientific evidence, it formulates demonstrably novel research hypotheses for experimental verification. The system’s design involves agents continuously generating, critiquing and refining hypotheses accelerated by scaling test-time compute. Key contributions include (1) a multi-agent architecture with an asynchronous task execution framework for flexible compute scaling, and (2) a tournament evolution process for self-improving hypotheses generation. Automated evaluations show continued benefits of test-time compute scaling, improving hypothesis quality over time. Although this is a general-purpose system, we focus the validation in three biomedical applications: drug repurposing; novel-target discovery 1 ; and explaining mechanisms of antimicrobial resistance 2 . Specifically, Co-Scientist helped to identify new drug-repurposing candidates and synergistic combination therapies for acute myeloid leukaemia that were validated through in vitro experiments. These real-world validations demonstrate the potential of Co-Scientist to accelerate scientific discovery and usher in an era of AI-empowered scientists.
1
Automated evaluations show that increasing test-time compute continually improves the quality of generated hypotheses over time.
2
Co-Scientist is a Gemini-based multi-agent AI system designed to generate demonstrably novel, experimentally testable scientific hypotheses from research objectives and prior evidence.
3
In biomedical validation tasks, Co-Scientist identified drug-repurposing candidates and synergistic combination therapies for acute myeloid leukaemia.
4
Its asynchronous multi-agent architecture enables flexible test-time compute scaling, while tournament evolution supports continuous critique, refinement, and self-improvement of hypotheses.
5
The identified acute myeloid leukaemia candidates and combination therapies were validated through in vitro experiments, demonstrating potential to accelerate scientific discovery.

Co-Scientist, a multi-agent AI system for scientific hypothesis generation, validated in biomedical applications

The system’s ability to generate, critique, refine, and improve the novelty and quality of experimentally testable scientific hypotheses, including drug-repurposing and synergistic-therapy hypotheses for acute myeloid leukaemia

Publication Details
Publication Date
2026-05-19
Journal
Publisher
ISSN
Cited by
11
Access Type
Author Information
Authors
Pushmeet Kohli
Demis Hassabis
Keran Rong
Nenad Tomašev
Gary Peltz
Tao Tu
Demis Hassabis
Ottavia Bertolli
Eeshit Dhaval Vaishnav
Juraj Gottweis
Yossi Matias
Alan Karthikesalingam
Vivek Natarajan
Katherine Chou
Khaled Saab
V. S. Dhillon
Guan Yuan
José R. Penadés
Alexander Daryin
Wei‐Hung Weng
Anatoly Myaskovsky
Annalisa Pawlosky
Tiago R. D. Costa
Vikram Dhillon
Tom Sheffer
Anil Palepu
Petar Sirkovic
Felix Weissenberger
Dan Popovici
Fan Zhang
Avinatan Hassidim
Amin Vahdat
K. Kulkarni
B. Lee
Yunhan Xu
Jan Freyberg
Juraj Gottweis
Wei-Hung Weng
Tao Tu
Petar Sirkovic
Grzegorz Głowaty
Felix Weissenberger
Alessio Orlandi
Dan Popovici
Anil Palepu
Ryutaro Tanno
Khaled Saab
Fan Zhang
Jacob Blum
Andrew Carroll
Kavita Kulkarni
Nenad Tomašev
Dina Zverinski
Ivor Rendulić
Elahe Vedadi
Florian Hasler
Luka Rimanić
Marina Boia
Ivan Budiselić
Ben Feinstein
Mathias Bellaiche
Jeremy Ratcliff
Katherine Chou
Avinatan Hassidim
Burak Göktürk
Amin Vahdat
Yuan Guan
Byron Lee
Tiago R. D. Costa
José R. Penadés
Gary Peltz
James Manyika
Yunhan Xu
Alan Karthikesalingam
Vivek Natarajan
Grzegorz Glowaty
Alessio Orlandi
Keran Rong
Ryutaro Tanno
Jacob Blum
Andrew Carroll
Dina Zverinski
Ivor Rendulic
Elahe Vedadi
Florian Hasler
Luka Rimanic
Marina Boia
Ivan Budiselic
Ben Feinstein
Mathias Bellaiche
Tom Sheffer
Jan Freyberg
Jeremy Ratcliff
Ottavia Bertolli
B. Gokturk
Eeshit Dhaval Vaishnav
James Manyika
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat
Make a presentation
100%