By: Krista Davidson
21 Jan, 2026
A Year of Impact for AI Safety in Canada
January 21, 2026
In 2025, the CAISI Research Program at CIFAR awarded $2.4M in funding to advance innovative research in AI safety
In just one year of the Canadian Artificial Intelligence Safety Institute (CAISI) Research Program at CIFAR, $2.4M was deployed to catalyze 12 groundbreaking research projects designed to identify and mitigate the risks of advanced artificial intelligence (AI) systems. The impact of this work is highlighted in the research program’s 2025 Year In Review Report: Building Safe AI for Canadians.
AI safety has emerged as a globally significant research field, dedicated to addressing the risks that AI poses to society. In response, the Government of Canada announced the launch of CAISI to enhance Canada’s capacity to address AI safety risks. Through CAISI, the government awarded CIFAR $27M to lead an independent, multidisciplinary program dedicated to advancing AI safety research.
The research program aims to address the sociotechnical challenges of AI safety, including algorithmic bias, discrimination and the spread of disinformation through deepfakes and AI-generated content; preventing cybersecurity attacks and models that provide inaccurate or unreliable information; and safeguarding against malfunctions, such as the loss of human control and oversight in AI systems as they become increasingly more autonomous.
Advancing High-Impact Sociotechnical Research
“Canada’s global reputation in AI is built on a foundation of research excellence and world-class talent. Through the CAISI Research Program at CIFAR, we are leveraging that strength and expertise to lead the next frontier and ensure that AI is not only powerful but also safe, trustworthy and aligned with human values,” says Elissa Strome, Executive Director of the Pan-Canadian AI Strategy at CIFAR.
The success of the CAISI Research Program lies in its ability to mobilize Canada’s exceptional AI talent necessary to understand and design ways to mitigate the risks of advanced AI systems.
“By fostering a critical mass of skilled talent and enabling multidisciplinary collaborations among the research community, government and industry, we are positioning Canada as a global leader in safe and trustworthy AI. This inaugural Year in Review report demonstrates that Canada is not only on the right track but also possesses the depth and caliber of expertise necessary to move the needle on the world’s most pressing AI safety challenges,” says CAISI Research Program Co-Directors Catherine Régis (Université de Montréal, Mila) and Nicolas Papernot (University of Toronto, Vector Institute).
The report highlights many extraordinary projects awarded in 2025, including the CIPHER project, which is developing a “human-in-the-loop” AI system designed to empower civil society organizations to detect and neutralize coordinated disinformation campaigns; Mila’s AI Safety Studio which is proactively addressing the risks of companion chatbots and creating safeguards to protect vulnerable users; and Vector Institute’s landmark study of 11 world-leading AI models, assessing their reliability and accuracy to ensure they are safe for high-stakes environments.
Read the CAISI Research Program at CIFAR’s 2025 Year in Review Report: Building Safe AI for Canadians. Access a full PDF of the report here.
Related Articles
-
Building safer AI with advanced evaluation methods
December 03, 2025
-
CIFAR launches new AI safety Networks to address synthetic evidence in the legal system and linguistic inequality
November 19, 2025
-
Calls open for global AI alignment research initiative
August 05, 2025
-
CIFAR announces first AI safety Catalyst Grants under new national program
June 04, 2025