Independent teams have converged on multi-agent, tool-using LLM architectures as the core design pattern for AI-for-science systems: Coscientist (Boiko et al. 2023) pairs GPT-4 with search, code execution and lab automation, Co-Scientist (Gottweis et al. 2026) is a multi-agent system built on Gemini, and the Virtual Lab (Swanson et al. 2024) has an LLM principal investigator guiding specialist agents. Qi et al. 2026's survey argues this convergence is functional, since multi-agent redundancy offers error-checking that can mitigate hallucination and bias.
E19, E5, and E44 confirm all three system descriptions (GPT-4 with search/code execution/lab automation; multi-agent Co-Scientist on Gemini; LLM principal investigator guiding specialist agents), and E24 states that multi-agent architectures offer structured redundancy and error-checking mechanisms that may help mitigate hallucination and bias.
Written by Kimi K3 via Ollama Cloud · checked by GLM-5.3 via Ollama Cloud · 30 Sept, 23:08
Source chain
AEvery quote below was checked, without a model, to appear verbatim in its source.
- 01
“an artificial intelligence system driven by GPT-4 that autonomously designs, plans and performs complex experiments by incorporating large language models empowered by tools”
Full-text passage · no page number · evidence E19
Passage read from www.ebi.ac.uk, which may be a preprint rather than the published version.
- 02
“a multi-agent artificial intelligence (AI) system built on Gemini for structured scientific thinking and hypothesis generation”
Full-text passage · no page number · evidence E5
Passage read from www.ebi.ac.uk, which may be a preprint rather than the published version.
- 03
“an LLM principal investigator agent guiding a team of LLM agents with different scientific backgrounds”
Abstract · no page number · evidence E44
- 04
“Multi-agent architectures, such as planners, executors, validators, and critics, offer structured redundancy and error-checking mechanisms”
Full-text passage · no page number · evidence E24
Passage read from www.ebi.ac.uk, which may be a preprint rather than the published version.