Artificial Intelligence and the Transformation of Scientific Rationality

The article examines the epistemic status of results produced by artificial intelligence (AI) systems in scientific inquiry when their internal logic remains methodologically opaque. It critically analyzes algorithmic empiricism, which declares theoretical hypotheses and explanatory models redundant in favor of large-scale automated detection of statistical correlations. Drawing on philosophy of science and the causal approach, the author argues that predictive accuracy neither amounts to causal explanation nor provides sufficient grounds for trusting conclusions, while data’s dependence on collection conditions necessitates substantive interpretation. Yet algorithmic opacity does not itself preclude a result’s scientific value. Examples from protein structure prediction, mathematical proof search, hypothesis generation, and autonomous laboratories demonstrate that results can be justified independently through formal verifiers or experiments that detect rather than reproduce search errors. The author distinguishes justification, contribution to understanding, and responsibility for use, recognizing the value of reliable predictions or proofs without equating them with explanations of the phenomena studied. A hybrid research model distributes cognitive operations between humans and machines, with reciprocal adjustments to data, algorithms, and theoretical propositions. Six criteria support its comprehensive evaluation: human relevance, methodological reconstructability, verifiability and contestability, traceability of AI contributions, increased understanding, and institutional responsibility with openness to criticism. The article concludes that AI integration does not overturn the ideals of rationality but complicates justificatory norms: extensive automation requires stronger independent verification, consensus on criteria for understanding both initial and resulting data, revised mechanisms for community recognition of discoveries, and unconditional researcher accountability for scientific conclusions.

1. Cory-Wright R., Cornelio C., Dash S., El Khadir B., & Horesh L. (2024) Evolving Scientific Discovery by Unifying Data and Background Knowledge with AI Hilbert. Nature Communications. Vol. 15, article 5922.

2. Duede E. (2023) Deep Learning Opacity in Scientific Discovery. Philosophy of Science. Vol. 90, no. 5, pp. 1089–1099.

3. Felin T. & Holweg M. (2024) Theory Is All You Need: AI, Human Cognition, and Causal Reasoning. Strategy Science. Vol. 9, no. 4, pp. 346–371.

4. Hempel C.G. (1998) The Logic of Explanation (O.A. Nazarova, Comp. & Trans.). Moscow: Dom intellektualnoy knigi; Russkoye fenomenologicheskoye obshchestvo (Russian translation).

5. Jumper J., Evans R., Pritzel A., Green T., Figurnov M., Ronneberger O., …, & Hassabis D. (2021) Highly Accurate Protein Structure Prediction with AlphaFold. Nature. Vol. 596, no. 7873, pp. 583–589.

6. Kapoor S. & Narayanan A. (2023) Leakage and the Reproducibility Crisis in Machine-Learning-Based Science. Patterns. Vol. 4, no. 9, article 100804.

7. Karniadakis G.E., Kevrekidis I.G., Lu L., Perdikaris P., Wang S., & Yang L. (2021) Physics-Informed Machine Learning. Nature Reviews Physics. Vol. 3, no. 6, pp. 422–440.

8. Kitchin R. (2014) The Data Revolution: Big Data, Open Data, Data Infrastructures and Their Consequences. London: SAGE Publications.

9. Kuhn T.S. (1975) The Structure of Scientific Revolutions (I.Z. Naletov, Trans.). Moscow: Progress (Russian translation).

10. Lakatos I. (2003) The Methodology of Research Programmes. Moscow: AST; Ermak (Russian translation).

11. Lektorsky V.A. (2025) Humanism versus Transhumanism: Prognosis and Project. Russian Journal of Philosophical Sciences = Filosofskie nauki. Vol. 68, no. 1, pp. 15–31 (in Russian).

12. Leonelli S. (2016) Data-Centric Biology: A Philosophical Study. Chicago: University of Chicago Press.

13. Lu C., Lu C., Lange R.T., Foerster J., Clune J., & Ha D. (2024) The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery. arXiv preprint. arXiv:2408.06292.

14. MacLeod B.P., Parlane F.G.L., Morrissey T.D., Häse F., Roch L.M., Dettelbach K.E., …, & Berlinguette C.P. (2020) Self-Driving Laboratory for Accelerated Discovery of Thin-Film Materials. Science Advances. Vol. 6, no. 20, article eaaz8867.

15. Merchant A., Batzner S., Schoenholz S.S., Aykol M., Cheon G., & Cubuk E.D. (2023) Scaling Deep Learning for Materials Discovery. Nature. Vol. 624, no. 7990, pp. 80–85.

16. Messeri L. & Crockett M.J. (2024) Artificial Intelligence and Illusions of Understanding in Scientific Research. Nature. Vol. 627, no. 8002, pp. 49–58.

17. Norelli M.F., Votsis I., & Williamson J. (2025) The Interplay of Data, Models, and Theories in Machine Learning. Philosophy of Science. Vol. 92, no. 5, pp. 1383–1393.

18. O’Neil C. (2018) Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy (V. Degtyareva, Trans.). Moscow: AST (Russian translation).

19. Pearl J. & Mackenzie D. (2023) The Book of Why: The New Science of Cause and Effect (T. Mamedova & M. Antipov, Trans.). Moscow: AST (Russian translation).

20. Popper K.R. (2004) The Logic of Scientific Discovery (V.N. Sadovsky, Ed. of Russian trans.). Moscow: Respublika (Russian translation).

21. Romera-Paredes B., Barekatain M., Novikov A., Balog M., Kumar M.P., Dupont E., …, & Fawzi A. (2024) Mathematical Discoveries from Program Search with Large Language Models. Nature. Vol. 625, no. 7995, pp. 468–475.

22. Rudin C. (2019) Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead. Nature Machine Intelligence. Vol. 1, pp. 206–215.

23. Schölkopf B., Locatello F., Bauer S., Ke N.R., Kalchbrenner N., Goyal A., & Bengio Y. (2021) Toward Causal Representation Learning. Proceedings of the IEEE. Vol. 109, no. 5, pp. 612–634.

24. Sergeev S.F. (2026) Synthetic Data Autophagy: The Ecological Limit of Generative Artificial Intelligence Evolution. Russian Journal of Philosophical Sciences = Filosofskie nauki. Vol. 69, no. 1, pp. 26–50 (in Russian).

25. Song P., Yang K., & Anandkumar A. (2024) Towards Large Language Models as Copilots for Theorem Proving in Lean. arXiv preprint. arXiv:2404.12534v1.

26. Sullivan E. (2024) Do Machine Learning Models Represent Their Targets? Philosophy of Science. Vol. 91, no. 5, pp. 1445–1455.

27. Symons J. & Alvarado R. (2022) Epistemic Injustice and Data Science Technologies. Synthese. Vol. 200, no. 2, article 87.

28. Termine A., Ratti E., & Facchini A. (2026) Machine Learning and Theory-Ladenness: A Phenomenological Account. Synthese. Vol. 207, no. 3, article 94.

29. Thurston W.P. (1994) On Proof and Progress in Mathematics. Bulletin of the American Mathematical Society. Vol. 30, no. 2, pp. 161–177.

30. Trinh T.H., Wu Y., Le Q.V., He H., & Luong T. (2024) Solving Olympiad Geometry without Human Demonstrations. Nature. Vol. 625, no. 7995, pp. 476–482.

31. Wang H., Fu T., Du Y., Gao W., Huang K., Liu Z., …, & Zitnik M. (2023) Scientific Discovery in the Age of Artificial Intelligence. Nature. Vol. 620, no. 7972, pp. 47–60.

32. Willard J., Jia X., Xu S., Steinbach M., & Kumar V. (2023) Integrating Scientific Knowledge with Machine Learning for Engineering and Environmental Systems. ACM Computing Surveys. Vol. 55, no. 4, article 66, pp. 1–37.