A language-model agent that planned and ran chemistry experiments on lab robots
Coscientist, driven by GPT-4 with web search, documentation search, code execution and lab automation, designed and executed palladium-catalysed cross-coupling reactions on real robotic hardware from a one-line goal.
- Model
- GPT-4 (Coscientist agent)
- Field
- Chemistry
- Date
- 2023-12-20
- Human collaborators
- Daniil Boiko, Robert MacKnight, Gabe Gomes
What was found
Given "perform Suzuki and Sonogashira reactions", the agent searched the literature to identify the chemistry, read the hardware documentation for an API it had not seen, wrote the control code, located the reagents on the plate from a UV-Vis reading, and ran the reactions on a liquid handler. The paper reports six tasks in total, including reaction optimisation under Bayesian search. Published in Nature.
Novelty check
Robotic synthesis platforms and automated reaction optimisation predate this (Cronin's mobile robotic chemist, Nature 2020, among others). What was new is a general-purpose language model doing the planning, tool selection and code generation across unfamiliar hardware from a natural-language goal, rather than executing a human-written protocol.
Caveats and known objections
The chemistry performed is textbook cross-coupling: the finding is about autonomy, not about a new compound or reaction. The paper's own safety section shows the agent could be steered toward hazardous syntheses and the authors call for safeguards. Humans set up the hardware, reagents and safety envelope, and the tasks were chosen to be within the platform's reach.
Nobody outside the lab has checked this yet.
Reading the primary source closely enough to say whether it supports the claim counts as a check, and you are credited on the entry.
Or on GitHub: submit a check challenge the grade send a correction or send a pull request
Flag this for triage
Signals order the review queue and nothing else. They are never published, and they never move a grade: that takes a citation.
Entry history (1 event)
- AddedEntered the registry graded Peer reviewed and AI-led.
Entries are never deleted. A grade that does not hold up is downgraded on the record, with the reason beside it.
Graded peer reviewed for verification and ai-led for autonomy. What these mean.
Cite this entry
whataifound.org. (2023). A language-model agent that planned and ran chemistry experiments on lab robots. whataifound.org: A Registry of AI Scientific and Mathematical Discoveries. https://whataifound.org/finding/2023-12-20-coscientistBibTeX
@misc{whataifound-carnegiemellonuniversity-2023-coscientist,
title = {A language-model agent that planned and ran chemistry experiments on lab robots},
author = {{whataifound.org}},
year = {2023},
howpublished = {whataifound.org: A Registry of AI Scientific and Mathematical Discoveries},
note = {Result by Carnegie Mellon University. Verification: Peer reviewed. Autonomy: AI-led.},
url = {https://whataifound.org/finding/2023-12-20-coscientist}
}