•  
  •  
 

Bulletin of Chinese Academy of Sciences (Chinese Version)

Keywords

artificial intelligence, research paradigm, scientific organization management, AI-native, innovation system

Abstract

Artificial intelligence (AI) is profoundly reshaping research paradigms. This study aims to analyze the intrinsic mechanisms through which AI empowers scientific research and its impact on the organizational management models of research. By summarizing what AI can and cannot do in empowering research, it reveals the current effectiveness and capability boundaries of AI in this domain. Based on the extraction of common core conditions for AI-empowered research and the deconstruction of typical cases of AI-enabled research organizational models, this study analyzes the differences between the organizational management model of AI-empowered research and traditional research organizational models. Furthermore, it proposes three major contradictions that need to be resolved in AI-empowered research: organized research vs. free exploration, platform-based supply vs. personalized demand, and intelligent research productivity vs. traditional research management systems. Finally, the study preliminarily proposes the mechanisms and pathways for building an AI-native scientific research innovation system and for promoting the transformation and innovation of research organizational management models.

First page

1403

Last Page

1415

Language

Chinese

Publisher

Bulletin of Chinese Academy of Sciences

References

[1] The Royal Society. Science in the age of AI: How artificial intelligence is changing the nature and method of research. London: The Royal Society, 2024.

[2] Tshitoyan V, Dagdelen J, Weston L, et al. Unsupervised word embeddings capture latent knowledge from materials science literature. Nature, 2019, 571: 95-98.

[3] Eger S, Cao Y, D’Souza J, et al. Transforming science with large language models: A survey on AI-assisted scientific discovery, experimentation, content generation, and evaluation. (2025-02-09)[2026-03-14]. https://doi.org/10.48550/arXiv.2502.05151.

[4] 李国杰. 智能化科研（AI4R）：第五科研范式. 中国科学院院刊, 2024, 39(1): 1-9. Li G J. AI4R: The fifth scientific research paradigm. Bulletin of Chinese Academy of Sciences, 2024, 39(1): 1-9. (in Chinese)

[5] 孙凝晖. AI4S的若干思考. 计算, 2025, 1(5): 16-19. Sun N H. Several thoughts on AI for science. Computing Magazine of the CCF, 2025, 1(5): 16-19. (in Chinese)

[6] Mirhoseini A, Goldie A, Yazgan M, et al. A graph placement methodology for fast chip design. Nature, 2021, 594: 207-212.

[7] Qiao K K, Chen A, Wang L D, et al. Optimizing the number of floorplanning layers for stacked integrated circuits based on spiking variational graph auto-encoders. Information Sciences, 2026, 724: 122681.

[8] Pham T D, Tanikanti A, Keçeli M. ChemGraph as an agentic framework for computational chemistry workflows. Communications Chemistry, 2026, 9: 33.

[9] Zhang D, Jia X, Tran H B, et al. “DIVE” into hydrogen storage materials discovery with AI agents. Chemical Science, 2026, 17(6): 3031-3042.

[10] Xie T, Fu X, Ganea O E, et al. Crystal diffusion variational autoencoder for periodic material generation. (2021-12-15)[2026-03-14]. https://openreview.net/forum?id=03RLpj-tc_.

[11] Zhang D, Bi H R, Dai F Z, et al. Pretraining of attention-based deep learning potential model for molecular simulation. NPJ Computational Materials, 2024, 10: 94.

[12] Zhang D, Liu X, Zhang X Y, et al. DPA-2: A large atomic model as a multi-task learner. NPJ Computational Materials, 2024, 10: 293.

[13] Jumper J, Evans R, Pritzel A, et al. Highly accurate protein structure prediction with AlphaFold. Nature, 2021, 596: 583-589.

[14] Zhang L F, Han J Q, Wang H, et al. Deep potential molecular dynamics: A scalable model with the accuracy of quantum mechanics. Physical Review Letters, 2018, 120(14): 143001.

[15] Zhang L F, Han J Q, Wang H, et al. End-to-end symmetry preserving inter-atomic potential energy model for finite and extended systems. Advances in Neural Information Processing Systems, 2018, 31: 6440-6449.

[16] Walmsley M, Lintott C, Géron T, et al. Galaxy Zoo DECaLS: Detailed visual morphology measurements from volunteers and deep learning for 314000 galaxies. Monthly Notices of the Royal Astronomical Society, 2021, 509(3): 3966-3988.

[17] Watson-Parris D. Machine learning for weather and climate are worlds apart. Philosophical Transactions Series A, Mathematical, Physical, and Engineering Sciences, 2021, 379(2194): 20200098.

[18] Feng T, Trinh T, Bingham G, et al. Semi-autonomous mathematics discovery with gemini: A case study on the Erdős problems. (2026-01-22)[2026-03-14]. https://arxiv.org/abs/2601.22401.

[19] Barreto K, Kang J, Kim S H, et al. Irrationality of rapidly converging series: A problem of Erdős and Graham. (2026-01-21)[2026-03-14]. https://arxiv.org/abs/2601.21442.

[20] Woodruff D P, Cohen-Addad V, Jain L, et al. Accelerating scientific research with gemini: Case studies and common techniques. (2026-02-07)[2026-03-14]. https://doi.org/10.48550/arXiv.2602.03837.

[21] Wu Y K, Zhao H Y, Arora S. Unrealized expectations: Comparing AI methods vs classical algorithms for maximum independent set. (2025-02-05)[2026-03-14]. https://doi.org/10.48550/arXiv.2502.03669.

[22] Liu Y, Zhou C, Chen Y C, et al. Hard constraints meet soft generation: Guaranteed feasibility for LLM-based combinatorial optimization. (2026-02-01)[2026-03-14]. https://doi.org/10.48550/arXiv.2602.01090.

[23] Coveney P, Highfield R. AI needs physics more than physics needs AI. (2025-12-18)[2026-03-14]. https://doi.org/10.48550/arXiv.2512.16344.

[24] Kent A. Fundamental physics, existential risks and human futures. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 2025, 383: 20230376.

[25] Raikov A. Architecture of full-analogue photonic AI for non-standard problem solving. Frontiers in Physics, 2026, 13: 1704910.

[26] Endres R G. The unreasonable likelihood of being: Origin of life, terraforming, and AI. (2025-07-24)[2026-03-14]. https://doi.org/10.48550/arXiv.2507.18545.

[27] Gu J Y, Zhang X, Wang G L. Beyond the norm: A survey of synthetic data generation for rare events. (2025-06-04)[2026-03-14]. https://doi.org/10.48550/arXiv.2506.06380.

[28] Lancelin A, Bouchet F, Wikner A, et al. Solving the lack of data issue for machine learning for rare climate events// Proceedings of the European Geosciences Union General Assembly 2026. Vienna: Copernicus Publications, 2026: EGU26-20131.

[29] Hao Q Y, Xu F L, Li Y, et al. Artificial intelligence tools expand scientists’ impact but contract science’s focus. Nature, 2026, 649: 1237-1243.

[30] Wu Q C, Li Z R, Han B C, et al. Wafer-scale ultrathin and uniform van der Waals ferroelectric oxide. Science, 2026, 391: eadz1655.

[31] 李国杰. 基于可判定性理论的人工智能系统安全风险分类. 计算机研究与发展, 2026, 63(3): 539-547. Li G J. A safety risk taxonomy of AI systems based on decidability theory. Journal of Computer Research and Development, 2026, 63(3): 539-547. (in Chinese)

Download
Request a Copy

Share

COinS