Krueger AI Safety Lab (KASL)
David Krueger's Research Group at Cambridge, formally known as the Krueger AI Safety Lab (KASL), was an AI safety research group based at the University of Cambridge's Computational and Biological Learning Lab (CBL) and Machine Learning Group (MLG) from 2021 to 2024. Led by David Krueger as Assistant Professor, the group conducted research on alignment failure modes, algorithmic manipulation, interpretability, robustness, reward gaming, goal misgeneralization, and understanding how AI systems learn and generalize. The lab published extensively at top venues including ICML, NeurIPS, and ICLR, and ran paid internship programs in collaboration with the ERA Fellowship. Krueger departed Cambridge in 2024 to join the University of Montreal and Mila, where the lab continues under the KASL name.
Funding Details
- Annual Budget
- -
- Monthly Burn Rate
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
- Fiscal Sponsor
- -
Theory of Change
The group's theory of change centered on the belief that reducing existential risk from AI requires both deep technical understanding of how AI systems fail and broader efforts in governance and coordination. By researching alignment failure modes such as goal misgeneralization, reward gaming, and algorithmic manipulation, the group aimed to identify and characterize the ways advanced AI systems could become dangerous before those failures occur in high-stakes deployments. Their work on interpretability and robustness sought to make AI systems more transparent and reliable. By publishing at top venues, training new AI safety researchers through PhD supervision and internship programs, and engaging in policy work such as the UK AI Safety Institute and the CAIS Statement on AI Risk, the group aimed to both advance the technical frontier of AI safety and build the human capital and institutional capacity needed to govern transformative AI.
Grants Received
from Survival and Flourishing Fund
Projects
No linked projects.
People
No linked people.
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.
Details
- Last Updated
- Apr 3, 2026, 1:16 AM UTC
- Created
- Mar 18, 2026, 11:18 PM UTC