Policy & governance institutes
Where the public conversation about how to regulate advanced AI is actually researched and written.
Most AI policy is drafted in three rooms: legislatures, agencies, and the think tanks that brief both. The institutes below produce the analysis that ends up in EU AI Act guidance, US executive orders, and OECD principles. Their public reports are the fastest route to understanding why a given regulation looks the way it does.
- 01
Stanford HAI
2019 · Stanford, USAStanford's Institute for Human-Centered AI. Publishes the annual AI Index — the most widely cited dataset on the state of the field.
PolicyAI Index - 02
MIT FutureTech
2018 · Cambridge, USAResearch group studying how compute, data, and algorithmic progress drive AI capability, with widely cited work on scaling and forecasting.
Compute trendsForecasting - 03
OECD.AI Policy Observatory
2020 · Paris, FranceCross-national hub for AI policy indicators, the OECD AI Principles, and the live tracker of national AI strategies.
International policyIndicators - 04
Centre for the Governance of AI (GovAI)
2018 · Oxford, UKIndependent research centre focused specifically on the governance of transformative AI and frontier-lab policy.
AGI governance - 05
CSET
2019 · Washington, DC, USAGeorgetown's Center for Security and Emerging Technology. Produces detailed analysis of AI talent, compute, and national-security implications.
SecurityCompute supply chains - 06
Brookings AI & Emerging Tech
1916 · Washington, DC, USAMainline US think-tank programme on AI regulation, labour-market effects, and democratic-society impacts.
US policyRegulation - 07
RAND Technology and Security
1948 · Santa Monica, USARAND's emerging-tech work spans AI in defence, biosecurity risks of advanced models, and frontier-AI governance.
National securityRisk - 08
AI Now Institute
2017 · New York, USAResearch institute focused on the social and concentrated-power implications of AI, including labour, surveillance, and competition policy.
Civil rightsPower - 09
Ada Lovelace Institute
2018 · London, UKUK independent institute studying the public-interest governance of data and AI, with influential work on UK foundation-model policy.
Public interestFoundation-model policy - 10
Center for a New American Security (CNAS)
2007 · Washington, DC, USAUS defence and national-security think-tank with a substantial AI programme on military applications, export controls, and US-China competition.
Defence AIGeopolitics
How to use this list: when a new AI rule is announced, search two or three of these institutes for prior reports on the same topic — the rule almost always rhymes with their recommendations.