The global conversation around artificial intelligence (AI) is complex, multifaceted and nuanced. It's not a single-issue discussion, but one that touches on a variety of interconnected areas with far-reaching implications. Amidst this continuous exploration, AI Trust and Safety remains a priority focus.
In November 2024, Canada made a bold investment in AI safety with the creation of the Canadian AI Safety Institute (CAISI). The launch signalled a clear recognition that trust and adoption depend on making safety a central part of AI’s development. It also positioned Canada within a growing International Network of Safety Institutes advancing AI safety for the world’s most advanced AI systems.
Trust & Safety at Amii
Amii is proud to be part of this leading effort. In January 2025, we welcomed Alyssa Lefaivre Škopac as our first Director of AI Trust & Safety to build a portfolio that advances Amii’s research excellence and broadens our impact in this critical field.
“What excites me most about this work is the chance to build on the breakthroughs and positive impacts I’ve already seen, while also taking the risks seriously. AI safety often gets framed as being in tension with innovation, as if we have to choose between advancing progress or keeping it safe. Both matter, and that’s the work we’re doing here. By moving past that false binary, we can build trust and make real progress.”
In addition to leading Amii’s Trust & Safety team, Alyssa Škopac represents Amii on the research council for the Canadian AI Safety Institute (CAISI) Research Program at CIFAR, alongside Amii Fellow and CIFAR AI Chair Geoffrey Rockwell. The program recently awarded its first AI Catalyst Grants to Amii Fellows and CIFAR AI Chairs Matthew Taylor, Mi Jung Park and Randy Goebel.
These projects build on the research strength of Amii’s Fellows—spanning technical areas such as explainability, fairness, evaluation, and privacy, as well as domain expertise in health, philosophy, robotics, and governance and beyond. Together, this breadth expands Canada’s capacity to better understand the risks and opportunities of advanced AI systems and to explore practical ways of addressing them.
AI safety often gets framed as being in tension with innovation, as if we have to choose between advancing progress or keeping it safe. Both matter, and that’s the work we’re doing here. By moving past that false binary, we can build trust and make real progress.
Alyssa Lefaivre Škopac
Director of AI Trust and Safety
Focused and Ambitious
Impact in this landscape requires clarity of focus and deliberate choices about where to lead.
“The technology, capabilities, and research landscape are moving so quickly. It’s actually quite crazy trying to keep up,” said Škopac. “We know we will need to adapt, evolve, and respond as things change. But impact means focus and making clear choices about where to begin.”
Amii’s leadership in reinforcement learning (RL) makes it the natural starting point for the Trust & Safety research program. As one of the most powerful techniques in modern AI, RL presents important considerations for safety and trust, which will only grow in significance as the technology matures. By focusing here first, Amii is building on deep expertise while looking ahead to the issues that will shape the global AI safety conversation. The aim is to ensure RL methods are not only advancing the science but are also robust, reliable, and ready for real-world adoption.
Looking Around the Corner
Amii’s Trust & Safety research builds on strong foundations in research, literacy, talent development, and ecosystem leadership. These strengths provide the ability to focus where impact is clearest, while also creating room to expand into new areas of inquiry.
Škopac emphasizes that while Amii’s Trust & Safety research is focused, it is also designed to adapt as the landscape shifts. Staying live and informed about what is emerging next is central to how the work is shaped.
“The discourse is messy, definitions shift, and the path forward is not always obvious,” she explains. “But progress comes from testing assumptions, working together, and moving forward even in the absence of perfect clarity.”
This work is not happening in a vacuum. Amii collaborates across its national and international networks, filling gaps, building connections, and contributing where its strengths add the most value. The approach is clear: stay focused, remain agile, and stay connected to the wider ecosystem shaping AI’s future.
Building Trust, Supporting Adoption
Trust is not a switch to be flipped. It is built over time through consistent action, alignment between words and practice, and collaboration across communities.
Amii’s Trust & Safety work is contributing to this long-term effort by advancing trust and safety research in reinforcement learning, engaging across disciplines, and working with partners to broaden impact and enable responsible adoption
Through this work, Amii is reinforcing Canada’s position as a global leader in shaping the future of AI trust and safety.
Škopac sums up: “It’s about building the foundations of trust that make adoption possible. That means being consistent, thoughtful, and forward-looking, while recognizing we aren’t doing this alone. This work is deeply connected and collaborative. It’s hard, but it’s really important. And that’s what makes it exciting.”
Learn more about Amii's Trust & Safety program and stay connected with the latest developments HERE.