Eric Joel Horvitz is an American computer scientist and a Technical Fellow at Microsoft, where he became the company's first Chief Scientific Officer in March 2020 (Source: en.wikipedia.org; engadget.com). He is a former president of the Association for the Advancement of Artificial Intelligence, and his contribution to AI policy is institutional as much as scholarly: he founded and endowed the standing long-horizon study of AI's societal effects that also produced the AI Index.
Roles and appointments
Horvitz has served on the President's Council of Advisors on Science and Technology, on the Scientific Advisory Committee of the Allen Institute for Artificial Intelligence, and on the Computer Science and Telecommunications Board of the US National Academies (Source: en.wikipedia.org). His 2008 AAAI Presidential Lecture, "Artificial Intelligence in the Open World," and his earlier work on continual computation predate the current policy cycle; the AAAI presidency is the position from which he commissioned the one-year study on long-term AI futures that became AI100.
One Hundred Year Study on Artificial Intelligence
In 2014 Horvitz defined and funded, with his wife, the One Hundred Year Study on Artificial Intelligence at Stanford University, a project committed to a perpetual series of studies by multidisciplinary panels assessing AI's immediate, longer-term and far-reaching effects. He has stated that the founding gift is sufficient to fund the study for a century. Its declared topics include abuses of AI that could threaten democracy and freedom, and the possibility of superintelligence and loss of control (Source: en.wikipedia.org; ai100.stanford.edu). The AI Index was launched in 2016 as a project of the study.
The study's first report, published in 2016 under panel chair Peter Stone, took a position on long-horizon risk that its own authors described as deliberate. The report stated that "unlike in the movies, there's no race of superhuman robots on the horizon or probably even possible," and Stone said it "was a conscious decision not to give credence to this in the report" — a stance that placed AI100's early output on the opposite side of the risk debate from the arguments collected at AI Existential Risk. Stone's own framing of labor effects was that AI would supplement rather than automatically replace human workers (Source: en.wikipedia.org).
Positions
Writing in April 2024 with Vincent Conitzer, Sheila McIlraith and Peter Stone as the founder and leaders of AI100, Horvitz set out ten recommendations for AI research, policy and practice, published in Communications of the ACM in June 2024. The document argues against the field's division into short-term and long-term camps — "despite the growing divide in the field between focusing on short-term versus long-term implications, we think both are of critical importance" — and frames the coming decades as potentially "a turning point for humanity, comparable to the industrial revolution." Its methodological emphasis is on advancing scientific understanding of neural models and their behaviors as capabilities emerge, and on addressing effects through technical, social and sociotechnical lenses drawing on engineering, social, behavioral and economic expertise (Source: arxiv.org).
With Tom Mitchell he co-authored a treatment of scientific progress in AI (Scientific Progress in Artificial Intelligence: History, Status, and Futures — Eric Horvitz and Tom M. Mitchell (2024)), which supplies historical context to contemporary capability measurement: the measurement side establishes what current systems do and the historical side establishes what comparable claims have previously implied. See History of AI.
Relationships
- related: Scientific Progress in Artificial Intelligence: History, Status, and Futures — Eric Horvitz and Tom M. Mitchell (2024) — co-author
- related: Tom Mitchell — co-author
- related: Microsoft — employer; first Chief Scientific Officer
- related: President's Council of Advisors on Science and Technology (PCAST) — advisory service
- related: Stanford HAI AI Index Report 2026 — the AI Index was launched as an AI100 project
- related: History of AI, AI Benchmarks and Evaluation
Sources
- (Source: en.wikipedia.org) — biography, appointments, AI100 founding and funding, 2016 report positions
- (Source: ai100.stanford.edu) — AI100 description and Horvitz's role
- (Source: arxiv.org) — Horvitz, Conitzer, McIlraith and Stone, "Now, Later, and Lasting: Ten Priorities for AI Research, Policy, and Practice," CACM June 2024
- Scientific Progress in Artificial Intelligence: History, Status, and Futures — Eric Horvitz and Tom M. Mitchell (2024) — co-authored treatment of scientific progress in AI