Canada’s Big Push on AI

Canadian Prime Minister Justin Trudeau recently joined the Canadian machine learning community to celebrate the country’s burgeoning AI ecosystem at the University of Toronto Rotman School of Management’s Creative Destruction Lab (CDL). Held in late October, the event attracted scientists and investors, mostly from central Canada and the US east coast.

“I think we all understand, certainly in this room, the way the world is going,” said Trudeau with regard to the wide-reaching changes artificial intelligence is bringing. “So let’s be part of it and help shape it, and let’s make sure we’re benefiting from the innovations in design and the applications and jobs.”

Trudeau also said that like China and the United States, Canada regards the development of artificial intelligence technology from a top-down strategic perspective. This March, Canada announced a CDN$125 million “Pan-Canada AI Strategy”; in September Trudeau shared a stage with Jack Ma at the Gateway ’17 event to welcome Alibaba’s venture into Canadian online shopping; and in October Trudeau personally called Amazon CEO Jeff Bezos to invite him to build his company’s new headquarters in Canada. The city of Toronto also hooked up with Alphabet’s urban innovation department Sidewalk Labs to redevelop 12 acres of prime lakefront land this year.

Government top-down policy support is vitally important, but the Canadian industry’s immediate task is providing AI based services and products that attract investors and sell in the market.

Despite its strong R&D capabilities, Canada’s native artificial intelligence business prospects remain uncertain. While many global AI companies are tackling big challenges in manufacturing and healthcare; most of Toronto’s artificial intelligence companies are more narrowly focused on areas like real estate, marketing, entertainment, and fintech.

Montreal startup Element AI helps businesses drive revenue through artificial intelligence. Co-founded by Yoshua Bengio, it received CDN$137.5 million in Round A financing this summer and is expected to become an artificial intelligence unicorn. However, additional Canadian AI unicorns have thus far failed to materialize.

One big positive for Canadian AI is the emerging northward exodus of companies and talents from Silicon Valley. AI legal search company Ross Intelligence is one example of this. Founded by Canadians, the company is shifting operations from California back to Canada. Founder Jimoh Ovbiagele recently returned to Silicon Valley to headhunt Chinese engineers, exploiting widespread industry fears of American visa issues in the wake of Donald Trump’s election.

A healthy AI ecosystem is built on regional R&D clusters, capital, and policy. The average time required to train a doctoral student is five to six years, and the PhD mentor has great influence on the direction of the student’s research, and even how and where to develop it after graduation. In the coming few years, Canadian universities will graduate about 800 PhD students, who will have the advantage of studying under renowned scientists like Geoff Hinton, Yoshua Bengio, and Richard Sutton.

Canada’s commitment to AI is clear. British-Canadian researcher Geoffrey Hinton and Spanish-born Uber ATG Technology Director Raquel Urtasun have both said they regard Canada as the most suitable country for the development of generic artificial intelligence research. However, many still question whether a country with a population of just 36 million (about the same as California) can realistically attract and retain top talents in the field.

For many AI academics, the answer has less to do with Canada’s population and more to do with professional ethics and national policies. In 1986 Hinton left the US to avoid military funding from DARPA (the US Defense Advanced Research Projects Agency). Over the past 60 years DARPA has invested heavily in Carnegie Mellon University, MIT, the California Institute of Technology and other STEM Universities, effectively pushing artificial intelligence researchers into the weapons arena. In 2015 the US military invested US$149 million in 30 research projects on automated weapons. This surge in lethal autonomous weapons research has been vigorously protested by scientists such as Stuart Russell at the University of California, Berkeley.

At the CDL’s Q&A session, panelists were asked a simple question: “Which country do you think will win the artificial intelligence contest?” New York University Professor Gary Marcus immediately replied “China 2030 will dominate the world,” while Silicon Valley investor Steve Jurvetson said “no doubt, China or the United States.” Speaking with modesty befitting Canadian stereotypes, University of Alberta Professor Richard Sutton answered: “My hope is that no country will dominate this technology.”

Journalist: Meghan Han, Michael Sarazen