“Everyone should have a voice in tech”: Machine learning researcher Colin Raffel explains his decision to move from North Carolina to Toronto

Who: Colin Raffel, 37, associate professor of computer science, University of Toronto, and associate research director at the Vector Institute Known for: Researching how to make machine learning more democratic Moved from: University of North Carolina at Chapel Hill in July of 2023

My fascination with machine learning started with music. I was playing the guitar and making electronic music, and I became interested in the devices behind it—pedals, synthesizers, things like that. Eventually, I started building music software, and some of the tools I wanted to create required machine-learning techniques. I dove into the field, which led me to a PhD at Columbia University, where I worked in a lab that blended music and machine learning. From there, I did a residency at Google, then became an assistant professor of computer science at the University of North Carolina. Related: Trump’s Loss, Toronto’s Gain—Meet the artists, professors, scientists and other luminaries ditching the US and moving north

By early 2023, I decided it was time for a change. My campus was located in Chapel Hill, a small college town, and my partner and I wanted to live somewhere bigger and more cosmopolitan. I applied to universities across the northeastern US and two in Canada: the University of Toronto and McGill. At that point, we weren’t necessarily planning to leave the US, but when the Canadian offers came in, it made us stop and think seriously about what life could look like elsewhere.

Three main factors pulled me toward Toronto. First, I knew that U of T was an excellent place to work, and I was also offered a position at the Vector Institute—a globally renowned AI research centre that has been home to leading figures such as Raquel Urtasun and Geoffrey Hinton. Professionally, it was an incredible opportunity.

Related: Why Geoffrey Hinton is sounding the alarm about AI

Second, I already knew and liked the city after visiting several times for conferences over the years. The culture and diversity felt like a great fit. And third—and this became the most important one—my partner and I were planning to grow our family. At that point, our daughter was three, and we knew we wanted to have more kids. That got us thinking about the kind of environment we wanted them to grow up in.

Even though this was before Trump was elected for his second term, there was already a lot of uncertainty. The political climate in the US had felt unstable for years. And while I remained hopeful that things in America would eventually improve, I also recognized that change is slow and, in many cases, is met with great resistance. From where we were standing, Canada seemed like an appealing alternative. Not a utopia by any means, but saner and better aligned with our values.

Related: The Magic Lab—How chemist and computer scientist Alán Aspuru-Guzik brought self-driving laboratories from Harvard to Toronto

What mattered most to me was raising my kids in a society that values people who are different. I want them to grow up in a place where they can explore their gender identity or sexuality, where they’re exposed to different perspectives and cultures, and where diversity is embraced, not feared. I also want them to live in a country with a real social safety net—where someone facing a serious health issue doesn’t have to choose between going into poverty to get treatment and skipping care to stay financially solvent.

In the US, the erosion of fundamental principles such as freedom and equality has real consequences, not just in policy but in people’s everyday lives. It’s not just what gets legislated; it’s what becomes normalized. A presidential executive order declaring that there are only two genders doesn’t just affect laws—it trickles down into things like bullying at school, into hate and alienation in communities across the country. I didn’t want my kids growing up in that sort of world. So we made the move to Toronto in July of 2023. Our second daughter was born here—she’s Canadian through and through.

Related: “Trans people are terrified”—This LGBTQ immigration lawyer has been inundated with inquiries from the US since Donald Trump’s election

At U of T, my lab focuses on machine learning, which is in an era of rapid growth: larger models trained on massive datasets are producing major breakthroughs in fields including natural language processing—think chatbots like ChatGPT—and health care. But this expanding scale also creates barriers. Training these models requires enormous computational resources that only the wealthiest companies can afford. That means the power to build and shape this technology is at risk of being concentrated in very few hands.

My team is working to change that by facilitating the development of machine learning models collaboratively and in a decentralized way. Our goal is to make machine learning more ­democratic—to ensure that individuals and smaller organizations can meaningfully contribute to its development. When a technology becomes as powerful and pervasive as machine learning is now, it shouldn’t be controlled by a tiny minority. Everyone should have a voice. That principle is more important than ever in a time when social and political forces are pushing the world toward consolidation and exclusion.

The Vector Institute plays a huge role in supporting this work. It offers shared resources, including ­computing infrastructure and funding, but more importantly, it creates a community. My PhD students moved with me to Toronto, and as soon as they arrived, they were meeting with other researchers and building collaborations. If someone had a question or needed help trouble­shooting, they could walk across the office and talk to an expert. That kind of serendipitous interaction is invaluable.

When I think about the future of my field—and of my family—I’m happy I chose to be here along with so many like-minded colleagues. Our work is about more than just models and data. Ultimately, it’s about who gets to participate, whose voices are heard and what kind of society we’re building together.