Last reviewed
Sources
No citations yet
Review status
Needs citations
Revision
v5 · 3,764 words
Add missing citations, update stale details, or suggest a clearer explanation.
The Stanford Institute for Human-Centered Artificial Intelligence (HAI) is a university-wide research institute at Stanford University, founded in 2019 and co-directed at launch by computer-science professor Fei-Fei Li and former Stanford provost John Etchemendy, with the mission to "advance AI research, education, policy and practice to improve the human condition."[1][3] HAI is best known for two outputs that anchor much of how the field measures and describes itself: the annual AI Index Report, described by Stanford as "one of the most credible and authoritative sources for data and insights on artificial intelligence,"[26] and the 2021 report "On the Opportunities and Risks of Foundation Models," produced by HAI's Center for Research on Foundation Models (CRFM), which introduced the now-standard term foundation model.[13]
HAI was formally launched on March 18, 2019 at a symposium that drew speakers including Microsoft co-founder Bill Gates, California Governor Gavin Newsom, Jeff Dean of Google, Demis Hassabis of DeepMind, Reid Hoffman of Greylock Partners, and Eric Horvitz of Microsoft Research.[1][2] It was designed to bridge engineering, the humanities, and the social sciences, reflecting the view of its founders that AI's effects on society demanded multidisciplinary stewardship. The institute launched with about 200 participating faculty drawn from all seven schools of Stanford University, and at launch announced an intention to hire at least 20 additional faculty, including ten junior fellows, across fields spanning the humanities, engineering, medicine, the arts, and the basic sciences.[2] HAI publicly set a fundraising goal of approximately one billion U.S. dollars, comparable to a similar effort announced earlier the same year at the Massachusetts Institute of Technology.[1] As of its 2024 annual report, the institute had distributed more than fifty million U.S. dollars in grants to over 400 affiliated faculty across the university.[4]
Beyond the AI Index and its foundation-models work, HAI has become an important convener of academic, government, and industry voices on the regulation of frontier models, on AI in scientific discovery, and on the diffusion of generative AI across the U.S. economy. In May 2026, Stanford announced that HAI would be restructured and merged with the Stanford Data Science initiative under a unified institute that retains the HAI name, with computer scientist James Landay assuming the role of Denning Director (a senior academic chair named for foundational donors Steve and Roberta Denning) and co-founder Fei-Fei Li transitioning to a new role as a university-wide Special Advisor on AI.[5][6]
| Attribute | Detail |
|---|---|
| Founded | March 18, 2019[1] |
| Location | Stanford University, Stanford, California, United States[1] |
| Co-founding directors | Fei-Fei Li and John Etchemendy[1][2] |
| Current Director (2026) | James Landay (Denning Director)[5][6] |
| Executive Director | Russell Wald[7] |
| Affiliated faculty | More than 400 across all seven Stanford schools[4][6] |
| Notable programs | AI Index Report; CRFM; RegLab; DAWN; HAI Policy Hub; RAISE Health[8][9][10][11] |
| Initial fundraising target | ~$1 billion[1] |
| Cumulative grant funding (by 2026) | More than $60 million[6] |
Stanford University announced the creation of the Institute for Human-Centered Artificial Intelligence in early 2019 after roughly two years of internal planning by a group of faculty led by Fei-Fei Li, who had returned to Stanford from a role as Chief Scientist of AI/ML at Google Cloud, and Etchemendy, a logician and former Stanford provost (1999-2017). Other founding faculty included Christopher Manning, the computational linguist who would later be named the institute's first Thomas M. Siebel Professor in Machine Learning, and James Landay, a professor of computer science with a research focus on human-computer interaction.[1][2][5]
The launch event, held in Stanford's Frances C. Arrillaga Alumni Center on March 18, 2019, framed HAI's mission around three pillars: research, policy and society, and education. Stanford President Marc Tessier-Lavigne told the audience that artificial intelligence had the potential to "radically change how we live our lives" and that Stanford intended to "shape that future by putting humanists and social scientists alongside people who are developing artificial intelligence."[2] The same announcement set out a fundraising target of approximately one billion U.S. dollars to support new endowed professorships, fellowships, grants, and a planned new building.[1]
The phrase "human-centered AI", the institute's organising concept, was elaborated by Li in numerous talks and op-eds as comprising three commitments: (1) AI should be informed by an understanding of human intelligence; (2) AI should augment rather than replace human capabilities; and (3) the development of AI should be guided by concern for its effects on individuals, communities, and society. Etchemendy emphasized that the second co-director's chair was intentionally awarded to a non-engineer, signalling that AI policy and ethics were to be treated as first-class questions alongside technical research.[2][3]
The institute's intellectual lineage runs through several earlier Stanford efforts. Stanford had been a centre of AI research since the founding of the Stanford Artificial Intelligence Laboratory (SAIL) in 1962. In 2014, Stanford launched the One Hundred Year Study on Artificial Intelligence (AI100), a long-horizon research initiative chaired by Eric Horvitz and originally hosted by the Stanford Center for the Study of Language and Information (CSLI). The AI Index Report grew out of AI100 in 2017 and became HAI's flagship publication after 2019. HAI itself can be read as an institutional consolidation of these earlier strands (SAIL on the technical side, AI100 on the long-horizon societal side, and CSLI on the philosophical and cognitive side) into a single university-wide institute.[15]
From 2019 onwards HAI operated as a university-level institute reporting jointly to the Stanford Provost and the Stanford Vice Provost and Dean of Research. Day-to-day leadership rested with the co-directors Fei-Fei Li and John Etchemendy, supported by a slate of Associate Directors drawn from across the university. The initial Associate Directors included Susan Athey (economics, Stanford University Graduate School of Business), Russ Altman (bioengineering and medicine), Christopher Manning (computer science and linguistics), James Landay (computer science and human-computer interaction), and Surya Ganguli (applied physics and neuroscience).[12]
The institute's advisory council, which played a central role in fundraising and external relations, was chaired at launch by Reid Hoffman of Greylock Partners. The founding council also included Jim Breyer (Breyer Capital), Jeff Dean (Google), Steve Denning (General Atlantic), John Hennessy (Stanford), Eric Horvitz (Microsoft Research), Bob King (Peninsula Capital), James Manyika (McKinsey), Marissa Mayer (Lumi Labs), Sam Palmisano (Center for Global Enterprise), Heidi Roizen (DFJ/Threshold Ventures), Eric Schmidt (Alphabet), Kevin Scott (Microsoft), Ram Shriram (Sherpalo Ventures), Vishal Sikka, Neil Shen (Sequoia), and Jerry Yang (AME Cloud Ventures).[1]
Subsequent leadership additions included Daniel E. Ho, the William Benjamin Scott and Luna M. Scott Professor of Law at Stanford Law School, who joined the HAI leadership team as an Associate Director and founding director of the Regulation, Evaluation, and Governance Lab (RegLab).[10] Russell Wald, who had served as HAI's Director of Policy and was a principal architect of the institute's National AI Research Resource advocacy, was elevated to Executive Director and given oversight of research, education, finance, communications, industry programs, and the policy and society hub.[7]
In May 2026, Stanford announced a major restructuring that consolidated HAI with the Stanford Data Science initiative into a single institute retaining the HAI name. Under the new arrangement, James Landay was appointed Denning Director, the senior endowed academic leadership chair named for foundational donors Steve and Roberta Denning. Co-founder Fei-Fei Li transitioned to a new university-wide role as Special Advisor on AI and co-chair (with former Stanford President John Hennessy) of the institute's advisory council. Hennessy was additionally named a Stanford HAI Special Advisor.[5][6]
HAI's research operations are organised around a portfolio of grants, fellowships, and embedded centers and labs. Many of the institute's signature programs are jointly affiliated with other Stanford units (e.g., the Stanford School of Engineering, Stanford Law School, the Freeman Spogli Institute for International Studies).
The Center for Research on Foundation Models (CRFM) is an interdisciplinary research center launched on August 18, 2021 as an HAI initiative dedicated to the study, development, and deployment of foundation models, large pretrained models adaptable to many downstream tasks.[8] CRFM was founded and is directed by Percy Liang, an associate professor of computer science at Stanford. The center hosted a public Workshop on Foundation Models on August 23-24, 2021, at which its inaugural report was released.[8][13]
CRFM's launch was accompanied by the publication of "On the Opportunities and Risks of Foundation Models", a roughly 200-page report led by Rishi Bommasani and co-authored by a total of 114 researchers from across Stanford. The report is credited with introducing the now-standard term "foundation model": its authors wrote, "We call these models foundation models to underscore their critically central yet incomplete character," and it remains one of the most cited multi-author surveys of the field.[13] At the center's launch, Percy Liang said foundation models "are radically changing how AI systems will be built."[8] Outputs of CRFM include the HELM benchmark for holistic evaluation of language models, the Mistral training-stability project (no relation to the company of the same name), and the Levanter and Marin projects on transparent model training.[13]
The RegLab is HAI's flagship policy laboratory, founded by Daniel E. Ho in approximately 2020 and operated jointly with Stanford Law School. RegLab partners with government agencies, including the U.S. Environmental Protection Agency, Internal Revenue Service, Department of Labor, and Santa Clara County, to prototype, deploy, and evaluate machine-learning tools for public administration. Ho, who serves as an HAI Associate Director, has held appointments on the National Artificial Intelligence Advisory Committee (NAIAC), as Senior Advisor on Responsible AI to the U.S. Department of Labor, and as a Public Member of the Administrative Conference of the United States.[10]
The Data Analytics for What's Next (DAWN) project was originally founded in 2017, before HAI's launch, by Stanford computer scientists Peter Bailis, Kunle Olukotun, Christopher Ré, and Matei Zaharia, and was subsequently affiliated with HAI as an infrastructure-oriented research center. DAWN focuses on systems and tools for end-to-end machine-learning application development. Notable DAWN systems include Snorkel (weak-supervision data labelling), DAWNBench (an end-to-end deep-learning benchmark), Weld (a common runtime for high-performance data analytics), and NoScope (efficient neural-network video queries).[11]
HAI's Policy and Society Hub, headed organisationally by Russell Wald, produces white papers, congressional testimony, and policy briefs aimed at federal and state lawmakers. It collaborates closely with the Stanford Cyber Policy Center at the Freeman Spogli Institute. Marietje Schaake, a former Member of the European Parliament (2009-2019), joined Stanford in 2019 as the inaugural International Policy Director of the Cyber Policy Center and the inaugural International Policy Fellow at HAI, a position from which she has written extensively on AI governance and digital sovereignty.[14]
Announced on June 14, 2023 as a partnership between Stanford Medicine and HAI, RAISE Health (Responsible AI for Safe and Equitable Health) was launched to define standards and safeguards for the use of AI in medicine. It is co-led by Lloyd Minor, then Dean of the Stanford School of Medicine, and Fei-Fei Li. The initiative held its inaugural symposium on May 14, 2024, and operates as a repository for AI standards, tools, models, and best practices in health care.[9]
HAI funds a substantial portfolio of fellowships and grants, including the HAI Junior Faculty Fellows program, the HAI Postdoctoral Fellowships, the HAI-Hoffman Yee Research Grants (named for Reid Hoffman and Michelle Yee), the HAI Google Cloud Credit Grants, and seed research grants administered through Stanford's seven schools. By the time of the 2026 restructuring, HAI had distributed more than sixty million U.S. dollars in cumulative grant funding.[4][6]
The AI Index Report is HAI's most-cited public output and one of the most widely consulted longitudinal surveys of the artificial-intelligence field. Stanford describes it as "one of the most credible and authoritative sources for data and insights on artificial intelligence," and its stated mission is to "provide unbiased, rigorously vetted, and globally sourced data for policymakers, researchers, journalists, executives, and the general public to develop a deeper understanding of the complex field of AI."[26] It compiles data on technical performance, research and development, economy and labour, education, policy and governance, ethics, and public opinion, and has been cited by policymakers in the United States, the United Kingdom, the European Union, and more than a dozen other governments.[26]
The Index was conceived by Stanford computer scientist Yoav Shoham and was launched publicly at the end of 2017 as an offshoot of the One Hundred Year Study on Artificial Intelligence (AI100), a long-running Stanford research initiative. Shoham assembled a founding steering committee that included Ray Perrault (SRI International), Erik Brynjolfsson (then at MIT, later Stanford), and Jack Clark (then OpenAI). Following HAI's 2019 launch, the Index was formally absorbed into the new institute and has been published annually under HAI's auspices ever since.[15]
The 2026 Index also highlighted a narrowing of the technical performance gap between leading U.S. and Chinese frontier systems, reporting that as of March 2026 the leading U.S. model led its closest Chinese competitor by just 2.7 %, and documented an approximately 89 % decline since 2017 in the net flow of AI researchers into the United States.[17][18]
HAI has been a sustained voice in U.S. federal AI policy. The institute's most consequential advocacy effort has been the campaign to establish the National AI Research Resource (NAIRR), a proposed federal program to provide academic and non-profit researchers with shared compute and curated datasets.
Following a roughly ten-month policy practicum jointly hosted by HAI and Stanford Law School, the institute published in late 2020 a white paper titled "Building a National AI Research Resource: A Blueprint for the National Research Cloud." Russell Wald, then HAI Director of Policy, was a principal author and lobbyist for the resulting legislation, which authorized the NAIRR Task Force convened by the National Science Foundation (NSF) and the White House Office of Science and Technology Policy (OSTP). HAI submitted a formal response to the Task Force in July 2022, and Stanford projects were among those greenlit in the subsequent NAIRR pilot.[19][20]
HAI faculty and staff have testified repeatedly before the U.S. Congress on topics including the regulation of frontier models, antitrust implications of AI markets, AI in the federal workforce, and AI's role in scientific research. Daniel Ho, for example, testified before the U.S. House Committee on Oversight and Accountability on "Governing Artificial Intelligence: Three Futures" in December 2023, drawing on RegLab's experience deploying AI tools inside federal agencies. Susan Athey advised the U.S. Congressional Budget Office on the economic impact of AI prior to taking her appointment as Chief Economist of the U.S. Department of Justice Antitrust Division in 2022. The institute publishes regular policy briefs and operates the Stanford Emerging Technology Review in partnership with the Hoover Institution.[10][21]
Through its joint work with the Stanford Cyber Policy Center, HAI contributes to international AI governance debates, notably through Marietje Schaake's engagement with the European Union's AI Act and her writings on digital sovereignty and platform regulation. HAI has also hosted dialogues with delegations from the United Kingdom AI Safety Institute, the European Commission, and the Singaporean and Indian governments, and several of its policy briefs have been cited in OECD, G7, and UN reports on AI governance.[14]
HAI's affiliated faculty roster, more than 400 strong by 2026, spans the full breadth of Stanford University and includes many of the most cited researchers in modern AI. Representative figures include:
Beyond Stanford-based scholars, HAI has hosted distinguished visiting fellows from policy and industry, including former senior officials from the U.S. National Security Council, the European Commission, and several U.S. cabinet departments.[4][22]
HAI is funded through a combination of endowed gifts, individual major donations, federal research grants, and corporate affiliate fees. The institute's launch was anchored by a foundational gift from Stanford trustee and General Atlantic chairman Steve Denning and his wife Roberta (for whom the senior academic directorship is named) and by additional gifts from members of the advisory council. Steve Denning, who chaired the Stanford Board of Trustees from 2012 to 2017 (Stanford's longest-serving board chair), died in April 2026; his obituary in Stanford Report cited HAI as among the major Stanford initiatives he helped seed.[23]
The Corporate Founding Members Program, launched alongside HAI in 2019, asks each founding member to make a three-year commitment of approximately five million U.S. dollars per year. Early corporate members included Accenture, JPMorgan Chase, McKinsey, Wells Fargo, and Google; the program subsequently expanded to include American Express, AXA, and others through a Corporate Affiliates tier with lower commitments.[24]
Cumulative grant disbursements grew steadily through the institute's first seven years. The 2024 HAI Annual Report indicated more than fifty million U.S. dollars in cumulative grants to over 400 faculty across all seven Stanford schools; by the time of the May 2026 restructuring, that figure had grown to more than sixty million dollars.[4][6]
On May 6, 2026, Stanford University announced that HAI and the Stanford Data Science initiative would be merged into a single university-wide institute retaining the HAI name. Under the new structure, James Landay, previously HAI Associate Director and a founding faculty member, was named the Denning Director of the combined institute. Fei-Fei Li transitioned to a new university-wide role as Special Advisor on AI and was named co-chair (with John Hennessy) of the institute's advisory council; Hennessy was also given the additional title of HAI Special Advisor. The merger brought together HAI's network of more than 400 scholars, its industry affiliates program, and roughly sixty million dollars in cumulative grants with Stanford Data Science's Marlowe high-performance computing cluster and its early-career scholar fellowship program.[5][6][25]
In interviews accompanying the announcement, Landay said the merged institute would press researchers to weigh impact on users, communities, and society from a project's inception through development, deployment, and maintenance.[5]