A leading Canadian research university founded in 1957, home to AI safety-relevant research programs including technical AI safety grants from Coefficient Giving and CIFAR's Canadian AI Safety Institute program.
A leading Canadian research university founded in 1957, home to AI safety-relevant research programs including technical AI safety grants from Coefficient Giving and CIFAR's Canadian AI Safety Institute program.
People– no linked people
Updated 05/18/26Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26The University of Waterloo was founded in 1957 in Waterloo, Ontario, Canada, initially as an engineering and science faculty. It has grown into one of Canada's premier research universities, with over 41,000 students, a global alumni network of 263,000 across 160 countries, and a reputation for co-operative education and entrepreneurship through programs like its Velocity startup incubator. Within the AI safety ecosystem, Waterloo's relevance stems primarily from specific research programs and grants. In December 2025, professors Pascal Poupart (Waterloo) and Sriram Ganapathi (Carleton) received a $412,500 grant from Coefficient Giving (formerly Open Philanthropy) for research on making agentic AI systems safer through constraint learning. Their work focuses on multi-agent orchestrators — systems that coordinate multiple AI agents — and aims to identify and prevent unsafe behaviors including privacy leaks, harmful content generation, and system failures. A key distinguishing feature of this research is its focus on unknown, implicit, or hard-to-specify unsafe behaviors, and on improving safety during training rather than only at deployment. Through Canada's CIFAR Canadian AI Safety Institute (CAISI) Research Program, Waterloo has additional AI safety engagement. Professor Maura Grossman co-directs a network addressing synthetic AI-generated content in the justice system ($700,000 over two years), and Professor Wenhu Chen is developing advanced evaluation methods for AI safety. The university's AI Group also hosts ongoing research on safe ML systems (Prof. Shai Ben-David), trustworthy AI (Prof. Robin Cohen), and AI law, ethics, and policy (Prof. Maura Grossman). The Waterloo Data and Artificial Intelligence Institute (Waterloo.AI), established in 2018, serves as the university-wide hub for AI research and collaboration, with an explicit emphasis on responsible innovation. The Cheriton School of Computer Science's AI Group focuses on ensuring AI systems are safe, fair, interpretable, ethical, and trustworthy — properties directly aligned with AI safety concerns. Waterloo's overall research enterprise is substantial: external research funding reached $231.2 million in FY2022, and the institution operates on an annual budget exceeding $1 billion CAD.
Theory of Change
Updated 05/18/26Waterloo's AI safety-relevant work operates through multiple channels: producing technical research on constraint learning and safe agentic systems that can be adopted by AI developers; developing evaluation benchmarks and methods that improve the field's ability to measure AI safety properties; training the next generation of researchers through graduate programs and co-op placements at AI companies; and informing policy through collaborations like the CIFAR-Canadian Public Policy journal partnership on AI safety. The underlying assumption is that improving technical methods for detecting, constraining, and evaluating unsafe AI behavior, and embedding those methods in training rather than only deployment, will reduce the risk of harmful AI systems being deployed at scale.
Grants Received
Updated 05/18/26Projects– no linked projects
Updated 05/18/26Discussion
No comments yet. Be the first to share your thoughts.