Agentic profiles for effective AI governance - Nature

Abstract The creation of effective governance mechanisms for artificial intelligence (AI) agents requires a deeper understanding of their core properties and the implications they have for deployment. This paper provides a characterization of AI agents that focuses on four dimensions: autonomy, efficacy, goal complexity and generality. We propose different gradations for each dimension and argue that each dimension raises unique questions about the design, operation and governance of these systems. Moreover, we draw on this framework to construct ‘agentic profiles’ for different kinds of AI agent. These profiles help to illuminate cross-cutting technical and non-technical governance challenges posed by different classes of AI agents, ranging from narrow task-specific assistants to highly autonomous general-purpose systems. By mapping out key axes of variation and continuity across four dimensions, agentic profiles provide developers, policymakers and members of the public with guidance for effective AI governance. This is a preview of subscription content, access via your institution Access options Access through your institution Access Nature and 54 other Nature Portfolio journals Get Nature+, our best-value online-access subscription 27,99 € / 30 days cancel any time Learn more Subscribe to this journal Receive 52 print issues and online access 185,98 € per year only 3,58 € per issue Learn more Buy this article Purchase on SpringerLink Instant access to the full article PDF. 39,95 € Prices may be subject to local taxes which are calculated during checkout Subjects Social sciences Scientific community References Feigenbaum, E. A. The art of artificial intelligence: themes and case studies of knowledge engineering. In Proc. 5th International Joint Conference on Artificial Intelligence (IJCAI ’77) Vol. 2, 1014–1029 (Morgan Kaufmann, 1977). Russell, S. J. & Norvig, P. Artificial Intelligence: A Modern Approach (Prentice Hall, 1995). Sutton, R. S. & Barto, A. Reinforcement Learning: An Introduction (MIT Press, 1998). Wooldridge, M. in Multiagent Systems: A Modern Approach to Distributed Artificial Intelligence (ed. Weiss, G.) 27–79 (MIT Press, 1999). Sumers, T. R., Yao, S., Narasimhan, K. & Griffiths, T. L. Cognitive architectures for language agents. In Trans. Machine Learning Research (2024). Yee, L., Chui, M. & Roberts, R. Why agents are the next frontier of generative AI. McKinsey & Company https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/why-agents-are-the-next-frontier-of-generative-ai/ (2024). Salesforce. Salesforce unveils Agentforce–what AI was meant to be. Salesforce https://www.salesforce.com/uk/news/press-releases/2024/09/12/agentforce-announcement/ (2024). Spataro, J. New autonomous agents scale your team like never before. The Official Microsoft Blog https://blogs.microsoft.com/blog/2024/10/21/new-autonomous-agents-scale-your-team-like-never-before/ (2024). Casper, S. et al. The AI agent index. Preprint at https://doi.org/10.48550/arXiv.2502.01635 (2025). Gabriel, I. et al. The ethics of advanced AI assistants. Preprint at https://doi.org/10.48550/arXiv.2404.16244 (2024). Kirk, H. R., Gabriel, I., Summerfield, C., Vidgen, B. & Hale, S. A. Why human–AI relationships need socioaffective alignment. Humanit. Soc. Sci. Commun. 12 , 728 (2025). Article Google Scholar Chan, A. et al. Harms from increasingly agentic algorithmic systems. In Proc. 2023 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’23) 651–666 (Association for Computing Machinery, 2023). Uuk, R. et al. A taxonomy of systemic risks from general-purpose AI. Preprint at https://doi.org/10.48550/arXiv.2412.07780 (2024). Kasirzadeh, A. Two types of AI existential risk: decisive and accumulative. Philos. Stud. 182 , 1975–2003 (2025). Article Google Scholar Weidinger, L. et al. Taxonomy of risks posed by language models. In Proc. 2022 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’22) 214–229 (Association for Computing Machinery, 2022). Bird, C., Ungless, E., & Kasirzadeh, A. Typology of risks of generative text-to-image models. In Proc. 2023 AAAI/ACM Conference on AI, Ethics, and Society (AIES ’23) 396–410 (Association for Computing Machinery, 2023). Hammond, L. et al. Multi-agent risks from advanced AI. Preprint at https://doi.org/10.48550/arXiv.2502.14143 (2025). Kolt, N. Governing AI agents. Notre Dame Law Rev. https://doi.org/10.2139/ssrn.4772956 (2025). Kraprayoon, J., Williams, Z., & Fayyaz, R. AI agent governance: a field guide. Preprint at https://doi.org/10.48550/arXiv.2505.21808 (2025). Chan, A. et al. Visibility into AI agents. In Proc. 2024 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’24) 958–973 (Association for Computing Machinery, 2024). Chan, A. et al. Infrastructure for AI agents. In Trans. Machine Learning Research (2025). SAE International. Taxonomy and definitions for terms related to driving automation systems for on-road motor vehicles. SAE International https://www.sae.org/standards/content/j3016_202104 (2021). Council of the European Union. Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence and amending Regulations (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act). Eur-LEX http://data.europa.eu/eli/reg/2024/1689/oj (2024). Tabassi, E. Artificial intelligence risk management framework (AI RMF 1.0). National Institute of Standards and Technology http://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf (2023). Goldman, A. I. A Theory of Human Action (Prentice Hall, 1970). Davidson, D. I. in Agent, Action, and Reason (eds Binkley, R. W., Bronaugh, R. N. & Marras, A.) 1–37 (Univ. Toronto Press, 1971). Dretske, F. Explaining Behavior: Reasons in a World of Causes (MIT Press, 1988). Dung, L. Understanding artificial agency. Philos. Q. 75 , 450–472 (2025). Article Google Scholar Ginet, C. On Action (Cambridge Univ. Press, 1990). O’Connor, T. Persons and Causes: The Metaphysics of Free Will (Oxford Univ. Press, 2002). Lowe, E. J. Personal Agency: The Metaphysics of Mind and Action (Oxford Univ. Press, 2008). Frankfurt, H. G. Freedom of the will and the concept of a person. J. Philos. 68 , 5–20 (1971). Article Google Scholar Taylor, C. in The Self: Psychological and Philosophical Issues (ed. Mischel, T.) 103–135 (Blackwell, 1977). Grossman, S. J., & Hart, O. D. in Foundations of Insurance Economics (eds Dionne, G. & Harrington, S. E.) 302–340 (Springer, 1992). Kauffman, S. & Clayton, P. On emergence, agency, and organization. Biol. Philos. 21 , 501–521 (2006). Article Google Scholar Bouvard, V. et al. A review of human carcinogens—part B: biological agents. Lancet Oncol. 10 , 321–322 (2009). Article PubMed Google Scholar Meincke, A. S. in Philosophy of Science (eds Christian, A., Hommen, D., Retzlaff, N. & Schurz, G.) 65–93 (Springer, 2018). Damşa, C. I., Kirschner, P. A., Andriessen, J. E. B., Erkens, G. & Sins, P. H. M. Shared epistemic agency: an empirical study of an emergent construct. J. Learn. Sci. 19 , 143–186 (2010). Article Google Scholar Elgin, C. Z. Epistemic agency. Theory Res. Educ. 11 , 135–152 (2013). Article Google Scholar Sosa, E. Epistemic agency. J. Philos. 110 , 585–605 (2013). Article Google Scholar Nickel, J. W. in Ethnicity and Group Rights (eds Shapiro, I. & Kymlicka, W.) 235–256 (New York Univ. Press, 1997). List, C. & Pettit, P. Group Agency: The Possibility, Design, and Status of Corporate Agents (Oxford Univ. Press, 2011). Tollefsen, D. Groups as Agents (Polity, 2015). Bratman, M. Shared Agency: A Planning Theory of Acting Together (Oxford Univ. Press, 2014). Shapiro, S. J. in Rational and Social Agency: The Philosophy of Michael Bratman (eds Vargas, M. & Yaffe, G.) 257–293 (Oxford Univ. Press, 2014). Le Besnerais, A., Moore, J. W., Berberian, B. & Grynszpan, O. Sense of agency in joint action: a critical review of we-agency. Front. Psychol. 15 , 1331084 (2024). Article PubMed PubMed Central Google Scholar Russell, S. J. Rationality and intelligence. Artif. Intell. 94 , 57–77 (1997). Article Google Scholar Rao, A. S. & Wooldridge, M. in Foundations of Rational Agency (eds Wooldridge, M. & Rao, A.) 1–10 (Springer, 1999). Silver, D. et al. Mastering the game of Go without human knowledge. Nature 550 , 354–359 (2017). Article ADS CAS PubMed Google Scholar Brown, N. & Sandholm, T. Superhuman AI for multiplayer poker. Science 365 , 885–890 (2019). Article ADS MathSciNet CAS PubMed Google Scholar Meta Fundamental AI Research Diplomacy Team (FAIR) et al. Human-level play in the game of Diplomacy by combining language models with strategic reasoning. Science 378 , 1067–1074 (2022). Article ADS MathSciNet Google Scholar Lake, B. M., Ullman, T. D., Tenenbaum, J. B. & Gershman, S. J. Building machines that learn and think like people. Behav. Brain Sci. 40 , e253 (2017). Article PubMed Google Scholar Park, J. S. et al. Generative agents: interactive simulacra of human behavior. In UIST '23: Proc. 36th Annual ACM Symposium on User Interface Software and Technology https://doi.org/10.1145/3586183.3606763 (ACM, 2023). Wang, X. & Zhou, D. Chain-of-thought reasoning without prompting. Adv. Neural Inf. Process. Syst. 37 , 66383–66409 (2024). Article Google Scholar Zelikman, E., Wu, Y., Mu, J. & Goodman, N. D. STaR: self-taught reasoner bootstrapping reasoning with reasoning. In Proc. 36th International Conference on Neural Information Processing Systems (NIPS’22) 15476–15488 (Curran Associates Inc., 2022). Anthropic. Claude 3.5 Sonnet Model Card Addendum https://www-cdn.anthropic.com/fed9cc193a14b84131812372d8d5857f8f304c52/Model_Card_Claude_3_Addendum.pdf (2024). Kavukcuoglu, K. Gemini 2.5: Our most intelligent AI model. Google https://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/ (2025). Jaech, A. et al. OpenAI o1 system card. Preprint at https://doi.org/10.48550/arXiv.2412.16720 (2024). Berlin, I. in Four Essays On Liberty 118–172 (Oxford Univ. Press, 1969). Dennett, D. C. Freedom Evolves (Viking, 2003). Vagia, M., Transeth, A. A. & Fjerdingen, S. A. A literature review on the levels of automation during the years. What are the different taxonomies that have been proposed? Appl. Ergon. 53 , 190–202 (2016). Article PubMed Google Scholar Klyubin, A. S., Polani, D. & Nehaniv, C. L. Empowerment: a universal agent-centric measure of control. In Proc. 2005 IEEE Congress on Evolutionary Computation 128–135 (IEEE, 2005). Chalmers, D. J. Reality+: Virtual Worlds and the Problems of Philosophy (Allen Lane, 2022). Spirtes, P., Glymour, C. N., & Scheines, R. Causation, Prediction, and Search (MIT Press, 2000). Pearl, J. Causality: Models, Reasoning, and Inference 2nd edn (Cambridge Univ. Press, 2009). Bareinboim, E., Correa, J. D., Ibeling, D., & Icard, T. in Probabilistic and Causal Inference: The Works of Judea Pearl (eds Geffner, H., Dechter, R. & Halpern, J. Y.) 507–556 (Association for Computing Machinery, 2022). Jaeger, J., Riedl, A., Djedovic, A., Vervaeke, J. & Walsh, D. Naturalizing relevance realization: why agency and cognition are fundamentally not computational. Front. Psychol. 15 , 1362658 (2024). Article PubMed PubMed Central Google Scholar Sacerdoti, E. D. Planning in a hierarchy of abstraction spaces. Artif. Intell. 5 , 115–135 (1974). Article Google Scholar Georgievski, I. & Aiello, M. HTN planning: overview, comparison, and beyond. Artif. Intell. 222 , 124–156 (2015). Article Google Scholar Kwa, T. et al. Measuring AI ability to complete long software tasks. Preprint at https://doi.org/10.48550/arXiv.2503.14499 (2026). Deb, K., Sindhya, K. & Hakanen, J. in Decision Sciences: Theory and Practice (eds Sengupta, R., Gupta, A. & Dutta, J.) 145–184 (CRC Press, 2016). Li, M. & Vitányi, P. An Introduction to Kolmogorov Complexity and Its Applications (Springer, 2019). Shannon, C. E. A mathematical theory of communication. Bell Syst. Tech. J 27 , 379–423 (1948). Article ADS MathSciNet Google Scholar Papadimitriou, C. H. in Encyclopedia of Computer Science (eds Ralston, A., Reilly, E. D. & Hemmendinger, D.) 260–265 (Wiley, 2003). Holland, J. H. in Complexity and Industrial Clusters (eds Curzio, A. Q. & Fortis, M.) 25–34 (Physica, 2002). Legg, S. & Hutter, M. Universal intelligence: a definition of machine intelligence. Minds Mach. 17 , 391–444 (2007). Article Google Scholar Goertzel, B. Artificial general intelligence: concept, state of the art, and future prospects. J. Artif. Gen. Intell. 5 , 1–48 (2014). Google Scholar Hutter, M. Universal Artificial Intelligence: Sequential Decisions Based on Algorithmic Probability (Springer, 2005). Morris, M. R. et al. Position: Levels of AGI for operationalizing progress on the path to AGI. In Proc. 41st International Conference on Machine Learning (ICML’24) 36308–36321 (JMLR.org, 2024). Ehtesham, A., Singh, A., Gupta, G. K., & Kumar, S. A survey of agent interoperability protocols: Model Context Protocol (MCP), Agent Communication Protocol (ACP), Agent-to-Agent Protocol (A2A), and Agent Network Protocol (ANP). Preprint at https://doi.org/10.48550/arXiv.2505.02279 (2025). Anderljung, M. et al. Frontier AI regulation: managing emerging risks to public safety. Preprint at https://doi.org/10.48550/arXiv.2307.03718 (2023). Kasirzadeh, A. Measurement Challenges in AI Catastrophic Risk Governance and Safety Frameworks (Tech Policy Press, 2024). Luo, Z., Kasirzadeh, A., & Shah, N. B. The more you automate, the less you see: hidden pitfalls of AI scientist systems. Preprint at https://doi.org/10.48550/arXiv.2509.08713 (2025). Bowman, S. R. et al. Measuring progress on scalable oversight for large language models. Preprint at https://doi.org/10.48550/arXiv.2211.03540 (2022). Carpenter, J. in Living with Robots: Emerging Issues on the Psychological and Social Implications of Robotics (eds Park, R. de Visser, E. J. & Rovira, E.) 75–90 (Academic, 2020). Bereska, L. & Gavves, E. Mechanistic interpretability for AI safety – a review. Preprint at https://doi.org/10.48550/arXiv.2404.14082 (2024). Hacker, P., Edwards, L. & Kasirzadeh, A. AI, digital platforms, and the new systemic risk. In Proc. 2026 ACM Conference on Fairness, Accountability, and Transparency (FAccT '26) https://doi.org/10.1145/3805689.3806506 (ACM, 2026). Frey, C. B. & Osborne, M. A. The future of employment: how susceptible are jobs to computerisation? Technol. Forecast. Soc. Change 114 , 254–280 (2017). Article Google Scholar Webb, M. The impact of artificial intelligence on the labor market. SSRN Scholarly Paper at https://doi.org/10.2139/ssrn.3482150 (2020). Tyler, S. et al. Use of artificial intelligence in triage in hospital emergency departments: a scoping review. Cureus 16 , e59906 (2024). PubMed PubMed Central Google Scholar Salge, C., Glackin, C., & Polani, D. in Guided Self-Organization: Inception (ed. Prokopenko, M.) 67–114 (Springer, 2014). Luccioni, S., Jernite, Y. & Strubell, E. In Proc. 2024 ACM Conference on Fairness, Accountability, and Transparency (FAccT ’24) 85–99 (Association for Computing Machinery, 2024). Franklin, S. & Graesser, A. in Intelligent Agents III Agent Theories, Architectures, and Languages (eds Müller, J. P., Wooldridge, M. J. & Jennings, N. R.) 21–35 (Springer, 1997). Shavit, Y., et al. Practices for governing agentic AI systems. Research Paper, OpenAI https://cdn.openai.com/papers/practices-for-governing-agentic-ai-systems.pdf (2023). Masterman, T., Besen, S., Sawtell, M. & Chao, A. The landscape of emerging AI agent architectures for reasoning, planning, and tool calling: a survey. Preprint at https://doi.org/10.48550/arXiv.2404.11584 (2024). Huang, X. et al. Understanding the planning of LLM agents: a survey. Preprint at https://doi.org/10.48550/arXiv.2402.02716 (2024). Li, J. et al. EvoCodeBench: an evolving code generation benchmark with domain-specific evaluations. Adv. Neural Inf. Process. Syst. 37 , 57619–57641 (2024). Article Google Scholar Sharkey, L. et al. Open problems in mechanistic interpretability. Preprint at https://doi.org/10.48550/arXiv.2501.16496 (2025). Download references Acknowledgements We thank D. Abel, A. Chan, T. Everitt, F. Diaz, M. Franklin, A. Harutyunyan, A. Leicht, N. Lundblad, S. Krier, A. Manzini, A. Smart, M. Morris and the participants in the Foundations of Cooperative AI Lab (FOCAL) lunch series for their valuable feedback. The research of A.K. was supported by the AI2050 programme at Schmidt Sciences (grant 24-66924). Author information Author notes These authors contributed equally: Atoosa Kasirzadeh, Iason Gabriel Authors and Affiliations Carnegie Mellon University, Pittsburgh, USA Atoosa Kasirzadeh Google DeepMind, London, UK Atoosa Kasirzadeh & Iason Gabriel Authors Atoosa Kasirzadeh View author publications Search author on: PubMed Google Scholar Iason Gabriel View author publications Search author on: PubMed Google Scholar Corresponding authors Correspondence to Atoosa Kasirzadeh or Iason Gabriel . Ethics declarations Competing interests Both authors are employees of Google DeepMind. Peer review Peer review information Nature thanks Urs Gasser, Fabienne Marco, Ida Momennejad and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Additional information Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Rights and permissions Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. Reprints and permissions About this article Cite this article Kasirzadeh, A., Gabriel, I. Agentic profiles for effective AI governance. Nature 656 , 320–328 (2026). https://doi.org/10.1038/s41586-026-10805-z Download citation Received : 14 January 2025 Accepted : 04 June 2026 Published : 12 August 2026 Version of record : 12 August 2026 Issue date : 13 August 2026 DOI : https://doi.org/10.1038/s41586-026-10805-z Share this article Anyone you share the following link with will be able to read this content: Get shareable link Sorry, a shareable link is not currently available for this article. Copy shareable link to clipboard Provided by the Springer Nature SharedIt content-sharing initiative