IVADO
Runs R10, AI Safety and Alignment, a dedicated research grouping co-led by Yoshua Bengio, Chris Pal and Dhanya Sridhar, whose stated goal is to "accelerate this much-needed research towards developing AI agents that are safe and reliable". Three axes: evaluating and mitigating LLM biases and unsafe behaviours; improving LLM interpretability; and promoting safe and robust behaviours in multi-agent settings with LLMs. R10 sits inside IVADO, Québec's largest interdisciplinary AI research consortium, led by Université de Montréal with Polytechnique, HEC, Laval and McGill under a "robust, reasoning and responsible AI" (R3AI) strategy.