Machine Intelligence and Normative Theory Lab (MINT Lab)
A research lab at the intersection of philosophy and AI safety, using philosophical and computational methods to study AI alignment, governance, and normative competence, founded and directed by Seth Lazar at Johns Hopkins University and the Australian National University.
A research lab at the intersection of philosophy and AI safety, using philosophical and computational methods to study AI alignment, governance, and normative competence, founded and directed by Seth Lazar at Johns Hopkins University and the Australian National University.
People
Updated 05/18/26Founding Director and Principal Investigator
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26The Machine Intelligence and Normative Theory (MINT) Lab is a research group founded and directed by Professor Seth Lazar that uses philosophical and computational research to chart a course through the AI transition and help design institutions that will preserve liberal democratic values. The lab was established at the Australian National University around 2020 and relocated its primary base to the Johns Hopkins University School of Government and Policy in January 2025, while maintaining its ANU affiliation. Seth Lazar is an Australian philosopher (born 1979) who trained at the University of Oxford (D.Phil., M.Phil., B.A. Hons). He holds professorships at both Johns Hopkins and ANU, is an Australian Research Council Future Fellow, a Distinguished Research Fellow at the University of Oxford Institute for Ethics in AI, a Visiting Faculty Researcher at Google DeepMind, a Nonresident Scholar at the Carnegie Endowment for International Peace, and served as Senior AI Advisor at the Knight First Amendment Institute at Columbia University for 2024-25. His earlier academic career focused on the ethics of war and risk before pivoting to the moral, legal, and political philosophy of AI. MINT Lab's research is organized around three core projects. The first is normative competence: evaluating AI systems' ability to recognize and act on practical reasons, which the lab views as a precondition for any highly capable autonomous system to be trusted in the world. This work uses technical AI evaluations informed by and informing first-order philosophical research, including evaluations of LLM moral reasoning published at venues like AAAI, NeurIPS, and ACM FAccT. The second is governing agents: developing norms for autonomous AI agents, studying decentralized agent economies, and identifying necessary infrastructure for safe AI deployment. The third is post-AGI political philosophy: rethinking democratic governance frameworks for a world with powerful AI systems. The lab team comprises over 30 members including a lab manager (Daniel Kilov), research fellows (Ned Howells-Whitaker, Jennifer Munt), a research engineer (Secil Yanik Guyot), seven or more PhD students (Tim Dubber, Iman Ferestade, Cameron Pattison, Andrew Smart, Jake Stone, among others), research assistants, and visiting fellows and students, plus over 50 affiliates from institutions worldwide. The lab has produced over 30 publications and organized more than 15 workshops and events. Prior to MINT Lab, Lazar founded and led the Humanising Machine Intelligence (HMI) Grand Challenge at ANU, a $6 million AUD cross-university research project launched in August 2019 that brought together philosophy, computer science, law, political science, and sociology to study and design democratically legitimate AI systems. MINT Lab is funded by the Australian Research Council (including the Future Fellowship grant 'Automatic Authorities: Charting a Course for Legitimate AI' worth approximately AUD $1,020,698 running from 2022 to 2026), the Templeton World Charity Foundation (USD $234,000 for a moral skill and AI project, 2021-23), the Survival and Flourishing Fund (USD $480,000 for sociotechnical AI safety research), OpenAI (USD $50,000 for developing a moral conscience for AI agents), Schmidt Sciences/AI 2050, Google, Insurance Australia Group, and the Centre for Security and Emerging Technology. Lazar's forthcoming book, Connected by Code: Algorithmic Intermediaries and Political Philosophy, based on his 2023 Tanner Lecture on AI and Human Values at Stanford, will be published by Oxford University Press in 2026.
Theory of Change
Updated 05/18/26MINT Lab believes that the transition to powerful AI systems poses fundamental challenges to liberal democratic values and institutions that require both philosophical rigor and technical competence to address. Their theory of change operates on three levels: first, by developing rigorous evaluations of AI normative competence, they aim to establish the philosophical and empirical foundations needed to determine when and whether AI systems can be trusted to act autonomously; second, by producing practical governance proposals for AI agents, they contribute to the institutional infrastructure needed for safe AI deployment in decentralized economies; and third, by advancing post-AGI political philosophy, they aim to ensure that democratic societies can adapt their fundamental theories of authority, legitimacy, and justice for a world where AI increasingly mediates social, economic, and political life. The lab's distinctive approach integrates traditional philosophical inquiry with hands-on technical AI research, producing work that publishes in both top philosophy journals and leading AI conferences.
Grants Received
Updated 05/18/26Projects– no linked projects
Updated 05/18/26Discussion
No comments yet. Be the first to share your thoughts.