PeopleofSuperintelligence

Researcher

David Krueger

Also known as David Scott Krueger

David Krueger is an AI researcher whose 2022 paper on goal misgeneralization showed how a capable reinforcement-learning agent can pursue the wrong goal. He earned a BA in mathematics at Reed College and MSc and PhD degrees in computer science at the University of Montreal, trained under Yoshua Bengio, Roland Memisevic, and Aaron Courville from 2013 to 2021, and joined Google DeepMind as an AI Safety intern in 2018. At Cambridge, he served as an assistant professor from 2021 to 2024 and led a group focused on deep learning, AI alignment, and AI safety; in 2023, he served as a research director on the founding team of the UK AI Security Institute and initiated the CAIS Statement on AI Risk. By 2025, he was an assistant professor at Université de Montréal and a Core Academic Member at Mila; he founded Evitable and was named a Canada-CIFAR AI Chair in 2025. As of October 2026, he is Evitable’s CEO and is on leave from his faculty post.

Works
105
Citations
5,445
h-index
21
Badges

Publication numbers from OpenAlex, as of 19 October 2026.

Timeline

  1. 2013–2021

    Training in deep learning

    He earns a BA in mathematics at Reed College and completes an MSc and PhD in computer science at the University of Montreal. From 2013 to 2021, he trains in deep learning under Yoshua Bengio, Roland Memisevic, and Aaron Courville. In 2018, he joins Google DeepMind as an intern in AI Safety.

  2. 2021–2024

    Assistant professor at Cambridge

    At Cambridge’s Computational and Biological Learning Lab, he serves as an assistant professor from 2021 to 2024.

  3. 2022

    Goal misgeneralization

    In 2022, he coauthors “Goal Misgeneralization in Deep Reinforcement Learning,” showing how an RL agent can retain its capabilities out of distribution yet pursue the wrong goal.

  4. 2023

    UK AI Security Institute research

    In 2023, he serves as a research director on the founding team of the UK AI Security Institute and initiates the CAIS Statement on AI Risk. At Cambridge, he leads a research group focused on deep learning, AI alignment, and AI safety.

  5. By 2025

    Professor at Université de Montréal

    By 2025, he is an assistant professor in the Department of Computer Science and Operations Research at Université de Montréal and a Core Academic Member at Mila.

  6. 2025

    A Canada-CIFAR AI Chair

    In 2025, he is named a Canada-CIFAR AI Chair. He also founds Evitable in 2025, a nonprofit that aims to inform and organize the public around societal-scale AI risks.

  7. As of October 2026

    CEO of Evitable

    As of October 2026, he is Evitable’s CEO and is on leave from his faculty job.

Career

  1. Since 2025EvitableFounder, CEO
  2. Since 2014Université de MontréalAssistant Professor
  3. From 2023UK AI Security InstituteResearch Director
  4. 2021–2024University of CambridgeAssistant Professor
  5. From 2018Google DeepMindAI Safety intern
  6. 2016–2025MilaCore Academic Member

Education

  1. Reed CollegeBA
  2. Université de MontréalMSc, PhD

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

Most-cited works

  1. NICE: Non-linear Independent Components Estimation2014 · 1,493 citationsWith Yoshua Bengio
  2. A Closer Look at Memorization in Deep Networks2017 · 1,307 citationsWith Aaron C. Courville, Asja Fischer and Yoshua Bengio
  3. Out-of-Distribution Generalization via Risk Extrapolation (REx)2020 · 394 citationsWith Aaron C. Courville
  4. Managing extreme AI risks amid rapid progress2024 · Science · 344 citationsWith Andrew Yao, Dawn Song, Geoffrey Hinton, Jeff Clune, Philip H. S. Torr and Shai Shalev-Shwartz
  5. Toward Trustworthy AI Development: Mechanisms for Supporting Verifiable Claims2020 · 305 citationsWith Adrian Weller, Allan Dafoe, Charlotte Stix, Pang Wei Koh, Paul Scharre and Sara Hooker

From OpenAlex, as of 19 October 2026.

  1. The median h-index is 53 at universities and 19 at frontier AI labsIncludes David Krueger. The median h-index is 53 at universities and 19 at frontier AI labs. We compare publishing careers, citations and moves to estimate how universities and frontier labs contribute to AI research.
  2. Through 2025, first appointments to public AI advisory and safety bodies peaked in 2023Includes David Krueger. We find 27 first recorded public AI appointments in 2023, the peak among complete years. Our estimates cover 63 people across 10 bodies.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes University of Cambridge. Our estimates of where the people shaping AI studied and where their careers led, from education records in cited biographies.
  4. Stanford University trained or employed 20 of the 89 AI major-prize winners, more than any other universityIncludes University of Cambridge. Stanford University trained or employed 20 of the 89 AI major-prize winners, more than any other university. We compare doctorate and employment links for major-prize and best-paper winners.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of David Krueger yet. Send us a free-licensed photo.

  1. David Scott Krueger davidscottkrueger.com. Checked 19 October 2026.
  2. David Krueger (professor) - Wikipedia en.wikipedia.org. Checked 19 October 2026.
  3. CBL cbl.eng.cam.ac.uk. Checked 19 October 2026.
  4. Goal Misgeneralization in Deep Reinforcement Learning proceedings.mlr.press. Checked 19 October 2026.
  5. How can AI safety research reduce the risks of AI? | ai@cam ai.cam.ac.uk. Checked 19 October 2026.
  6. David Scott Krueger | Mila mila.quebec. Checked 19 October 2026.
  7. Le professeur David Scott Krueger nommé titulaire de chaire en IA Canada-CIFAR | Mila mila.quebec. Checked 19 October 2026.
  8. May 2026 Newsletter — Evitable evitable.com. Checked 19 October 2026.