The future of fundamental science led by generative closed-loop artificial intelligence
Review article, 2026
Artificial intelligence is approaching the point at which it can complete the scientific cycle, from hypothesis generation to experimental design and validation, within a closed loop that requires little human intervention. Yet, the loop is not fully autonomous: humans still curate data, set hyperparameters, adjudicate interpretability, and decide what counts as a satisfactory explanation. As models scale, they begin to explore regions of hypothesis and solution space that are inaccessible to human reasoning because they are too intricate or alien to our intuitions. Scientists may soon rely on AI strategies they do not fully understand, trusting goals and empirical payoffs rather than derivations. This prospect forces a choice about how much control to relinquish to accelerate discovery while keeping outputs human relevant. The answer cannot be a blanket policy to deploy LLMs or any single paradigm everywhere. It demands principled matching of methods to domains, hybrid causal and neurosymbolic scaffolds around generative models, and governance that preserves plurality and counters recursive bias. Otherwise, recursive training and uncritical reuse risk model collapse in AI and an epistemic collapse in science, as statistical inertia amplifies flaws and narrows the investigation. We argue for graded autonomy in AI-conducted science: systems that can close the loop at machine speed, while remaining anchored to human priorities, verifiable mechanisms, and domain-appropriate forms of understanding.
epistemic singularity
closed-loop discovery
graded autonomy
AI-conducted science
cognitive collapse
human-machine collaboration
domain-method alignment
AI4Science
Author
Hector Zenil
King's College London
The Francis Crick Institute
University of Cambridge
Alan Turing Institute
University of Oxford
Jesper Tegnér
King Abdullah University of Science and Technology (KAUST)
Karolinska Institutet
Felipe S. Abrahão
Laboratorio Nacional de Computacao Cientifica (LNCC)
State University of Campinas
King's College London
University of Oxford
Alexander Lavin
Pasteur Labs
Vipin Kumar
University of Minnesota
Jeremy G. Frey
University of Southampton
Adrian Weller
University of Cambridge
Alan Turing Institute
Larisa N. Soldatova
Goldsmiths, University of London
A. Bundy
University of Edinburgh
Nicholas R. Jennings
Loughborough University
Koichi Takahashi
RIKEN
Keio University
Lawrence Hunter
Colorado School of Public Health
Saso Dzeroski
Jozef Stefan Institute
Andrew Briggs
University of Oxford
Frederick D. Gregory
DEVCOM U.S. Army Combat Capabilities Development Command
Carla P. Gomes
Cornell University
Jon Rowe
University of Birmingham
Alan Turing Institute
James A. Evans
University of Chicago
Hiroaki Kitano
Okinawa Institute of Science and Technology Graduate University
Alan Turing Institute
Ross King
Alan Turing Institute
Chalmers, Computer Science and Engineering (Chalmers), Data Science and AI
University of Cambridge
Frontiers in Artificial Intelligence
26248212 (eISSN)
Vol. 9 1678539Subject Categories (SSIF 2025)
Philosophy
DOI
10.3389/frai.2026.1678539
PubMed
41755911