With the Canadian AI Safety Institute (CAISI) Research Program at CIFAR publishing its 2025 Year End Review: Building Safe AI for Canadians, Amii is proud to highlight our collaborative progress as a key partner. Alyssa Lefaivre Škopac, Amii's Director of Trust & Safety, shares a few of the milestones our team achieved in tandem with the CAISI Research Program.

2025 was about building the foundations of the Trust and Safety program at Amii. Following the launch of the Canadian AI Safety Institute in 2024, we focused on building the team, capacity, and partnerships needed to contribute meaningful science and research to some of the hardest technical and social questions around advanced AI systems.

We established a five-person team including program leadership, two Applied Research Scientists, a Machine Learning Engineer, and a postdoctoral fellow. In parallel, we shaped a focused and ambitious research agenda grounded in Amii's core strength in reinforcement learning. The work has been about setting direction, aligning on priorities, and putting the right people and structures in place to study vulnerabilities in frontier models and explore how new safeguards can be designed, tested, and evaluated.

Research is only one part of the picture. Our team spent considerable time in conversation with companies and executives about trust, safety, and real-world deployment, including at our flagship event, Upper Bound. Those conversations reinforced what we see repeatedly in practice: trust, safety, and innovation advance together.

Looking ahead to 2026, Trust and Safety at Amii is moving into execution. New research projects are underway, partnerships are deepening, and we're focused on producing work that shapes how frontier AI systems are built and deployed. With strong foundations in place, we're positioned to continue to deliver. Stay tuned. So much more to come!