Levels of Autonomy in Science Automation
DOI:
https://doi.org/10.51094/jxiv.818Keywords:
Artificial intelligence, AI-driven science, Automation of scienceAbstract
Artificial Intelligence (AI) in scientific and technological research is anticipated to be one of the fields with the most significant ripple effects among long-term AI applications. When AI is viewed as a form of automation technology, its utility is proportional to the extent to which it can reduce or simplify human operation and instruction. Consequently, the efficacy of scientific AI is considered to be closely related to its autonomy. This paper proposes a seven-level autonomy scale, ranging from 0 to 6, for scientific AI, primarily focusing on experimental science fields. This proposal draws inspiration from the autonomy levels for automated driving systems as defined by the Society of Automotive Engineers (SAE). This article discusses the definition of each level, the milestones for their realization, and their potential impacts on academia, technology, and society.
Conflicts of Interest Disclosure
The authors have no COI to disclose.Downloads *Displays the aggregated results up to the previous day.
References
Workshop summary, AI Scientist Grand Challenge, The Alan Turing Institute, London, February 26-27 (2020) https://www.turing.ac.uk/sites/default/files/2021-02/summary_of_discussion_workshop_2020_ai_scientist_grand_challenge_clean.pdf
高橋恒一, 科学AIの自律性レベル設定 Setting Levels for the Autonomy of Scientific AI, 人工知能学会第二種研究会資料, SIG-AGI-020-04 (2022)
岡村 麻子, 林 和弘, セミナーシリーズ「AIとデータで変わる科学と社会」理化学研究所 高橋 恒一 氏講演録：－AIは科学の営みをどう変えうるか－, STI Horizon 8(4) (2022)
Takahashi, K., Accelerating Life Sciences by Robotic Biology, poster presentation at AAAI Spring Symposium on Computational Approaches to Scientific Discovery (SS-23-04), March 27-29, Palo Alto, CA (2023)
渡部匡己, 都築拓, 海津一成, 高橋恒一, 「人工知能による科学研究の加速」, 人工知能学会全国大会論文集 2016, 30:1–4 (2016)
高橋恒一, 渡部匡己, 「現代科学を超えて―AI駆動型科学へ」 実験医学別冊 「あなたのラボにAI（人工知能）×ロボットがやってくる」, Eds. 夏目徹, 羊土社 (2017)
高橋恒一, 「第五の科学 自動化」In AI事典 第3版. 近代科学社 (2019)
Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles, SAE J3016 (2021)
Lindsay, R.K. et al., Applications of Artificial Intelligence for Organic Chemistry: The Dendral Project, McGraw-Hill Book Company (1980)
Shortliffe, E.H. and Buchanan, B.G. "A model of inexact reasoning in medicine", Mathematical Biosciences. 23 (3–4): 351–379 (1975)
Burger B. et al., A mobile robotic chemist, Nature 583(7815):237-241 (2020)
Kanda G. et al., Robotic search for optimal cell culture in regenerative medicine, eLife 11 (2022).
Coutant, A., et al., Closed-loop cycles of experimental design, execution, and learning accelerates systems biology model development in yeast, PNAS 116(36) (2019)
Lu, Chris et al. The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery. arXiv:2408.06292 (2024)
Bostrom, N., Superintelligence: Paths, Dangers, Strategies, Oxford University Press (2014)
Kraemer, R. D., and I. G. Harrison. A decision support system for the Military Airlift Command, the Airlift Deployment Analysis System. No. ORNL/TM-11201. Oak Ridge National Lab., TN (USA) (1989)
Popper, K.. The Logic of Scientific Discovery. Hutchinson & Co. (1959)
Taniguchi, T. et al, Symbol emergence in robotics: a survey, Advanced Robotics 30(11-12) (2016)
Udrescu, S., Tegmark, M., AI Feynman: a Physics-Inspired Method for Symbolic Regression, arXiv:1905.1148 (2019)
Schmidt M., Lipson H., Distilling Free-Form Natural Laws from Experimental Data, Science 324, 81 (2009)
Friston, K., The free-energy principle: a unified brain theory? Nature reviews neuroscience 11(2) (2010)
Taniguchi, T. et al, Collective Predictive Coding as Model of Science: Formalizing Scientific Activities Towards Generative Science, arXiv:2409.00102 (2024)
高橋恒一, 将来の機械知性に関するシナリオと分岐点, 人工知能 33(6) p. 867-872 (2018)
Takahashi, K., Scenarios and branch points to future machine intelligence, arXiv:2302.14478 (2023)
BEATLESS、長谷敏司、角川書店 (2012)
Chiang, T. Catching crumbs from the table. Nature 405, 517 (2000)
Downloads
Posted
Submitted: 2024-10-08 05:35:25 UTC
Published: 2024-10-15 09:20:52 UTC
License
Copyright (c) 2024
Koichi Takahashi
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
