The Stanford Institute for Human-Centered Artificial Intelligence (Stanford HAI) is an academic research institute at Stanford University working at the intersection of AI technology and policy. It is known for the annual AI Index Report, which tracks global AI trends across research, industry, policy, and public perception.
Overview
Stanford HAI is an academic research institute affiliated with Stanford University, founded in 2019 to advance AI research, education, policy, and practice with what it terms a human-centered framework. Its founders are Fei-Fei Li (co-director and the institute's public face), John Etchemendy (co-director and former Stanford provost), computer scientist Chris Manning, and James Landay (Source: hai.stanford.edu). By its fifth year the institute reported channeling more than $40 million into human-centered AI research supporting over 300 Stanford scholars across disciplines (Source: hai.stanford.edu).
It houses the Stanford Center for Research on Foundation Models (CRFM), a sub-center launched in 2021 that coined the term "foundation models," developed the HELM evaluation framework, and publishes the Foundation Model Transparency Index (FMTI); and the Stanford Digital Economy Lab (DEL), started in 2020 under Erik Brynjolfsson to study AI's effects on work, productivity, jobs, and inequality. HAI's affiliated researchers include co-director Li Fei-Fei, senior fellow Erik Brynjolfsson, Alex Pentland, CRFM director Percy Liang, and Cyber Policy Center co-director Nathaniel Persily.
Activities
AI Index Report
The AI Index Report 2026 is the ninth annual edition, tracking AI progress globally across R&D, performance, economy, science, medicine, education, policy, and public opinion. The 2026 report runs 423 pages and is organized around 15 top takeaways (Stanford HAI AI Index Report 2026).
Among the figures reported in the 2026 edition: US private AI investment reached $285.9B, roughly 23× China's $12.4B. Organizational adoption stood at 88%, with population adoption at 53% within three years. The US-China model gap was 2.7%, with Anthropic narrowly ahead as of March 2026. Employment of young developers (ages 22–25) declined approximately 20% from 2024. Reported productivity gains ranged from 14–26% in customer support and software development. Documented AI incidents totaled 362, up from 233 in 2024.
Policy engagement
HAI's policy work centers on informing AI governance and regulation. Its faculty have testified before Senate and House committees, met with agencies including the Department of Commerce and the Federal Trade Commission, and run an intensive policy boot camp for congressional staff (Source: hai.stanford.edu). HAI leaders, including deputy director Russell Wald, were early champions of a National AI Research Resource (NAIRR) to give academia and nonprofits access to compute and data concentrated in industry; they organized a coalition of universities and companies in 2020, published a blueprint in 2021, and the effort fed into the 2023 CREATE AI Act and a 2025 NSF pilot. RAISE-Health (Responsible AI for Safe and Equitable Health), launched with Stanford Medicine in June 2023 and co-led by Fei-Fei Li and medical-school dean Lloyd Minor, addresses AI safety and ethics in clinical care.
State-capacity and regulatory-burden research
On July 23, 2026, Stanford HAI and Stanford RegLab published findings from an AI system that scanned 500 million words of state statutes across all 50 states to identify reporting requirements, commissions and fees. The work appears in a paper by Daniel E. Ho, Emily Robitschek, Ananya Karthik, Gabe Malek and Derek Ouyang, "The Abundance of Reports and Incapacity of States," forthcoming in the Yale Journal on Regulation. Reporting requirements in California grew by 400% from 2000 to 2025, and 30% of ongoing California reports may never have been completed. Maryland agencies identified 20% of their reports as candidates for elimination or consolidation, and reading the state's reports was calculated to take up to 14 weeks against a 13-week legislative session. A single report was calculated to consume 3,500 staff hours and more than $870,000 to produce (Source: hai.stanford.edu).
The tool has fed state deregulatory efforts. New York Governor Kathy Hochul issued Executive Order No. 61 in July 2026 directing state agencies to carry out a regulatory reset; Zoe Jacobs, director of regulatory reform and delivery in Hochul's office, said the RegLab tool converted "unwieldy legalese into digestible datasets." Earlier work with the San Francisco City Attorney led to legislation streamlining over a third of the city's reporting requirements (Source: hai.stanford.edu). See California State Government (AI Deployer).
Other research
A study on organizational factors in corporate AI project success and failure identifies jurisdictional clarity, task centrality, and task enactment homogeneity as key variables (Karunakaran, Vendraminelli, Narayanan) (Source: hai.stanford.edu).
Relationships
- includes: Stanford Center for Research on Foundation Models (CRFM) — Stanford Center for Research on Foundation Models (sub-center; coined "foundation models"; publishes FMTI).
- home-of: Li Fei-Fei (co-director), Erik Brynjolfsson (senior fellow), Alex Pentland, Nathaniel Persily (Cyber Policy Center co-director).
- related: Rishi Bommasani — CRFM Society Lead.
- related: The Digitalist Papers (Stanford, Volumes 1–2) — Brynjolfsson/Pentland/Persily co-edited; published out of Stanford HAI.
- related: The California Report on Frontier AI Policy — Fei-Fei Li as Working Group co-lead; CRFM lead-writer role.