Science Utopia? Closed-Loop LLM Simulation of Academic Research Ecosystems
Abstract
Scientific progress emerges from a longitudinal ecosystem in which researchers, institutions, funding agencies, collaboration networks, and the scientific literature co-evolve. As AI becomes increasingly involved throughout the scientific research cycle, understanding these interconnected and evolving processes becomes increasingly important. We introduce SciUtopia, a persistent, closed-loop LLM-agent simulation framework for studying academic research ecosystems. SciUtopia models interconnected scientific processes such as research-direction choice, collaboration, submission, peer review, resubmission, citation, funding, and researcher attrition, while maintaining evolving states across simulated years. Its configurable institutional mechanisms and information channels provide a controlled testbed for matched counterfactual experiments and targeted interventions. Across 61 simulation worlds, SciUtopia simulates over 40,000 researchers from 8,000 institutions, producing around 400,000 publication decisions and 1.2 million LLM-generated peer reviews. Using these longitudinal simulations, we find that rejection-driven resubmission substantially amplifies reviewer burden beyond population growth alone, cautious exploration balances citation impact with career success and long-term topic diversity, and resource inequality can emerge even without detectable cumulative advantage from narrowly winning early funding. Code is available at https://github.com/Ahren09/ScienceUtopia.
Community
Top conferences now have over 30,000 submissions. What's next?
AI is accelerating research. How can our institutions keep pace? š¬
As the research community grows, peer review, funding, and career incentives help shape which ideas advanceāand who can continue pursuing them.
Our new paper, āScience Utopia? Closed-Loop LLM Simulation of Academic Research Ecosystems,ā introduces SUTO: a persistent, large-scale simulation of a scientific community, where thousands of AI researchers navigate evolving institutions, collaboration networks, and a growing body of literature.
Research choices, peer review, funding, and career trajectories unfold together over simulated decadesāwith each decision shaping the opportunities and constraints agents encounter next. This creates a testbed for studying how individual behaviors and institutional rules interact to produce long-term, system-wide outcomes.
Across 61 simulated worlds with 40,000+ researcher agents, we found some interesting patterns:
š Allowing more submissions increased publication output, but also reduced the share of researchers who remained active over timeāraising questions about how to support both productivity and sustained participation.
š§ Cautious exploration beyond existing expertise offered a promising balance between citation impact and career success, while helping sustain topic diversity.
āļø Narrowly winning an early grant did not necessarily translate into a lasting funding edge.
As AI expands what individual researchers can do, we hope SUTO can help us ask what a thriving scientific community needs.
Which rule or incentive in academia would you test first?
Get this paper in your agent:
hf papers read 2610.01257 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0
No model linking this paper
Datasets citing this paper 0
No dataset linking this paper
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper