"To make a real impact, we need to become more international"

Dr. Strome, you have a black belt in the martial art of taekwondo. Speed and quick reflexes are key in that sport—do these skills help you in your field of research, artificial intelligence?

Taekwondo requires two things: discipline and focus. I go to practice three times a week. You have to be disciplined if you want to achieve results. That’s undoubtedly a parallel to advancing research and working with AI. The second parallel is the intense focus required in this sport and in AI—the ability to concentrate, recognize clear opportunities, and block out background noise. And in AI, there’s a lot of noise.

Taekwondo is a contact sport. Do you also see the field of AI as a battle arena?

It’s an interesting analogy, no question—especially given the current race around AI. But in the realm of research, I don’t like the metaphor of battle and competition. We’re not interested in building the strongest, largest, or most comprehensive AI models; that shouldn’t be an end in itself.

So what is your focus?

Our goal is to truly understand these systems. Of course, we also want to develop and advance the technology—but in a responsible, strategic way that brings benefits to society. We want to balance technological progress by simultaneously developing an understanding of the risks and mitigating them. This brings us back to our focus: we want to harness the benefits without losing sight of the other challenges.

You yourself have been with CIFAR for nine years, working on the delivery of the Pan-Canadian AI strategy. What has happened since then?

When I started, universities were still largely working in silos, without much interaction with one another. Since then, we’ve built an interconnected national network anchored by world-class talent from coast to coast. Our three national AI institutes – Amii in Edmonton, Mila in Montreal and the Vector Institute in Toronto – are each building distinct AI ecosystems within their respective regions, but they collaborate seamlessly as part of a unified Canadian strategy.

Building on these achievements, what additional benefits do you expect from international collaborations, such as the EU’s initiative for an “Alliance for the Future” with Canada?

The challenges are too great for any one country to tackle alone. There are many areas where the EU’s strengths can complement our own, including AI. Germany is a standout example: it has a massive industry with a long tradition of research and development. AI has the potential to make an enormous difference there. In Canada, we have a smaller industrial R&D footprint, but deep foundational AI research. That makes international collaboration a great opportunity for both nations.

Could such collaboration also set a precedent in other research areas beyond AI?

Let’s take a specific look at Helmholtz: There are already quite a few successful collaborations between CIFAR and various Helmholtz Institutes. At CIFAR, we strive to bring the best people together and build interdisciplinary, international research networks. Helmholtz is highly attractive to talent from all over the world. Research excellence is therefore something that unites us. And both of our organizations have recognized how important international collaboration is—that’s a good sign!

Could you be more specific? Which fields are of particular interest for collaborations between Helmholtz and Canada?

There’s already a well-established connection in the neurosciences. Data science is another field, and these two areas, in turn, overlap significantly with AI research, which we’ve already discussed. Here’s another example: We’re currently preparing a collaboration with CISPA in the field of cybersecurity. And there’s undoubtedly potential in other fields as well. Here at CIFAR, we have more than a dozen different programs; we’re particularly strong in biomedicine, materials science, and the field of quantum materials—I’m convinced there are good opportunities for collaboration there.

Are international research collaborations more important today than they were ten or fifteen years ago?

Fifteen years ago, I was at the University of Toronto, and even back then, international partnerships were a priority. Collaboration in science has always helped us find better answers and better train the next generation of researchers. But of course, the demands on infrastructure are growing ever greater with new technologies—for that reason alone, it’s important to have strong partners. However, it’s also true that international partnerships are among the greatest challenges: You can’t just pair up two researchers and say, “Now go work together!” We have to find the right people—people who are a good fit, whose expertise complements each other, and who share similar goals and interests. That's exactly what we're good at here at CIFAR.

Isn’t the concept of geostrategic interests missing from your answer?

You’re absolutely right: Geopolitics has changed dramatically over the past 15 years. Canada has a long, multi-generational history of deep cooperation with the U.S., but we are also actively diversifying our global partnerships. Our Prime Minister, Mark Carney, often talks about how middle powers with shared values should join forces. And he speaks of “variable geometries”—meaning that it’s no longer about exclusive cooperation, but rather about flexible alliances in which we work with different partners on different issues.

Where do you see international research collaborations in 20 years?

I am convinced that future research networks with real impact must be even more international. Canada is now participating in the Horizon Europe programme, and that aligns well with our strategic focus on cooperation with Europe. I’m convinced that we should expand the programme and involve more Canadian researchers—such mechanisms for research funding and collaboration are absolutely essential. And I see another important area: the exchange of students and postdocs. Here at CIFAR, for example, we have a partnership with the ELLIS network…

…the European Laboratory for Learning and Intelligent Systems…

…and many early-career researchers are jointly mentored by professors in Canada and Europe. In doing so, they build personal relationships and get to know the ecosystem on both sides, with all its various facets. Programs like these are incredibly important for building the international collaborations of the future.