AI Scientists as Engines of Discovery: A Case for Development within Reformed Institutions
2026-06-22 • Artificial Intelligence
Artificial Intelligence
AI summaryⓘ
The authors explain that AI systems are starting to help with scientific discovery by doing tasks like reading papers, writing code, analyzing data, and suggesting new ideas. They say this change is a big shift, not just small improvements, and that future AI might act like 'AI scientists' working on their own. To make this work well and safely, the authors emphasize that scientific institutions need to change how they check and manage AI contributions. They also discuss new challenges for things like credit, peer review, and how humans will work with AI in science.
Agentic AIScientific discoveryMulti-agent systemsHypothesis generationModel criticismAccountabilityPeer reviewScience governanceAI safetyInterpretability
Authors
Raul Jimenez, Boris Bolliet, Francisco Villaescusa-Navarro, Rabih Zbib, Benjamin Wandelt, David N. Spergel, Thomas Meier, Jessica Montgomery, Hana Aliee, Licia Verde
Abstract
Agentic artificial intelligence (AI) systems are beginning to assist, accelerate, and partially automate scientific discovery, performing tasks that span literature synthesis, code generation, data analysis, hypothesis proposal, and model criticism. We argue that this transition is qualitative rather than incremental, and that suitably designed multi-agent systems may evolve from passive computational tools into ``AI scientists'' that can expand the hypothesis-generating and verification capacity of science. Such systems must be developed and deployed within a scientific ecosystem fit for purpose: institutions must be redesigned for verification, accountability, interpretability, and dual-use safety. We sketch how multi-agent architectures, illustrated by the prototype framework \textit{Denario}, accelerate the discovery cycle and traverse model spaces beyond human reach; examine what this implies for authorship, peer review, and the enduring role of human scientists; and close with recommendations for governing AI as an epistemic actor rather than a mere instrument.