- The Canadian government’s support for fundamental research, Canada’s great universities and a growing AI ecosystem have made Canada a world leader in AI.
- AI will transform every industry and sector of the Canadian economy, including energy, healthcare and financial services, and it will lead to healthier, more sustainable and more fulfilling lives for Canadians.
- We must prepare for the inevitable transformation in work that will come from algorithms performing many of the mundane tasks humans do now, and must develop the skillsets required to adapt to that change.
Many more established Canadian businesses must build their own AI capabilities or partner with Canadian AI companies to stay relevant in Canada’s future economy.
What is AI’s transformative potential for Canada’s future economy?
The world over, people think of Canada when they think of AI. The reality did not meet the hype for 50 years, until the deep learning technologies that were invented in Canada by Geoff Hinton, Yoshua Bengio and Rich Sutton came into the picture. Canada’s long-term commitment to curiosity-based research allowed individuals to continue to research a field that a lot of other computer scientists around the world felt was a dead-end and fringe science.
It is remarkable that all of these technologies only showed their applicability about 6 years ago. First with the Google Android speech-recognition system and later with the image recognition challenge, called ImageNet. Computer scientists have been trying to build this technology for 50 years, but it is only in the last few years that algorithms’ capabilities to perform useful tasks have developed.
“Canada’s long-term commitment to curiosity-based research allowed individuals to continue to research a field that a lot of other computer scientists around the world felt was a dead-end and fringe science.”
The first applications of deep learning were in perceptual tasks like speech, image and video recognition. Historical statistical and rule-based approaches were not at all suitable or flexible enough to perform these tasks. The first use of such technologies was in consumer tech companies in Silicon Valley, for example, Google, Amazon, Apple and Facebook. But now the second phase is underway, in which AI will impact all industries and sectors. The energy sector is definitely going to be transformed by AI, but so is the healthcare industry. Canada’s universal public health system enables us to have comprehensive digital health records, which are essential for machine learning. We also have a phenomenal medical research community. So, there will be innovative and transformational uses of AI in the healthcare sector and Canada can truly lead the world in this aspect.
In the long term, AI will help us to manage our resources more efficiently. For instance, a fully autonomous highway, that is, a highway with only self-driving cars, can move 3 times more traffic than a manually-driven highway. That is not only going to be a welcome change to anybody who commutes in and around Toronto or Montreal, but it is also a highly sustainable choice.
“There will be innovative and transformational uses of AI in the healthcare sector and Canada can truly lead the world in this aspect.”
As individuals, we will live much healthier lives. We will be able to prescribe personalized medicine and predict potential health outcomes, understand our genes and genetic makeup, and detect early signs of cancer and other health deficiencies, all with the help of algorithms. Moreover, we will live more sustainably and will not have to do as many mentally mundane jobs as we do today. It is wonderful that Canada can lead this transformation of our society into the future.
What are the potential risks that AI presents and how can we mitigate them in Canada?
There are two risks to AI: firstly, there is a bias built into AI models; and secondly, AI will impact jobs in the long-term. As data scientists, it is important for us to acknowledge and discuss these risks and to proactively manage their impact.
On the bias front, machine-learned algorithms are trained on historical data, so if there is bias in the data, it will also be reflected in the algorithm’s predictions. In order to mitigate that risk, we need to ensure that the data itself is accurate and that data scientists are very thoughtful and careful in data selection and algorithm training. Furthermore, AI systems cannot operate as black boxes; the AI models need to be explainable and testable for unfair bias.
Coming to AI’s long-term impact on jobs, all technological changes have impacted our labor market. For example, the advent of internal combustion engines in cars devastated the horse and buggy industry. That change is inevitable, and humans are highly adaptable to it. The net benefit of AI will be positive; we will live more prosperous and healthier lives. But we will also have to accept the necessity for a change in the skillsets required to work in a world where algorithms and computers take over a lot of mundane tasks that humans have performed historically. This change is coming, regardless of what Canada does.
Will Canada’s AI ecosystem be able to keep up with behemoths like the US or China in the long-term?
China and the United States are in a race for AI supremacy, and we need to be able to keep up with that. As long as individuals, companies and nations are competing with each other, there will be a tremendous amount of investment into AI. Canadian leadership in AI is a testament to the foresight of the country’s policymakers. It is remarkable that they made significant bets to support not only academic AI research, but also the commercialization of those efforts.
“China and the United States are in a race for AI supremacy, and we need to be able to keep up with that.”
Canada’s commitment to long-term education and academic research, and the government’s historical interest in AI are paying off. Canada’s long-term funding vehicles attracted individuals like Rich Sutton and Geoff Hinton who are leading AI research today. So, Canada’s advantage lies in a great post-secondary education system with amazing researchers.
The latest AI supercluster initiative is only the most recent example of federal support for AI in Canada. Through 2017’s budget funding, the federal government invested in an AI strategy as well as in AI institutes in Montreal, Toronto and Edmonton. Prior to that, it also provided funding to the University of Montreal for AI research. These funds amount to hundreds of millions of dollars and they are often doubled up by funds from provincial governments and from industry. For example, the Vector Institute in Toronto has received $230 million in funding; Ontario allocated $50 million to the institute, Ottawa allocated another $50 million and the remainder was funded by industry players. So, the federal government’s strategy on AI investments has had a catalytic effect to leverage multiple levels of government and industry to come together to make a bet on AI.
“The federal government’s strategy on AI investments has had a catalytic effect to leverage multiple levels of government and industry to come together to make a bet on AI. Having said that, Canada’s AI funding pales in comparison to China’s.”
Having said that, Canada’s AI funding pales in comparison to China’s. While there are hundreds of millions of dollars at play in Canada, there are billions of dollars at play in China. In addition, China has millions of more graduates entering the market with AI training. But when you compare Canada to other Western countries like the United States or even European countries, our AI investment-to-GDP ratio is significant. Lastly, in today’s geopolitical environment, Canada represents an amazingly attractive place for top researchers to come and practice their art.
How can AI impact the finance and banking sector, and why did you choose to keep Layer 6 in Canada?
Layer 6 was developed as an AI enterprise, taking some of the most advanced machine learning and deep learning technologies to large enterprises. At Layer 6, we found that banks had a tremendous amount of data but were very poor at converting it into value. That is what Layer 6 helped its clients with. We were pursued by a number of companies and chose to land at TD Bank. TD is not only Canada’s largest retail bank, but also a bank that very aggressively competes for and acquires market share in the United States. Out of all the Canadian companies that we considered, TD had the most data on customers. There is so much insight that our algorithms can generate using that data, which could help TD better anticipate customer needs and better service them.
“Canada needs local people to fund and start a company, build a talented organization, build intellectual property, and exit to a Canadian enterprise.”
Another motivation for selling to TD is that it helps us build up Canada’s economy. The domestic sale helps us provide value and bring new capabilities to an iconic Canadian brand. Prior to Layer 6, I had started previous tech companies, which I had sold to Yahoo and Microsoft. In the Layer 6 deal, unlike my two previous exits, my team and I did not have to move to the US. Moreover, it closes a full circle of what Canada’s entrepreneurial ecosystem needs. Canada needs local people to fund and start a company, build a talented organization, build intellectual property, and exit to a Canadian enterprise.
How do you see Canada’s AI ecosystem evolving in the coming years?
I prioritize helping the local AI ecosystem. I am the founder of the Toronto-based Vector Institute and one of the founding fellows of Toronto’s Creative Destruction Lab, which is the world’s largest AI accelerator. I hope that a lot more established Canadian enterprises will take AI seriously, and either look to invest in their own capabilities, or look to partner with Canadian AI companies.
“Today, a venture capital firm will strategically invest in Canadian companies and ask them to stay in Canada. In fact, they are asking their portfolio companies in the US to move to Canada to establish AI research labs here.”
The AI ecosystem has completely transformed in the last 5 years. Five years ago, most AI researchers in Canada tended to leave the country for Silicon Valley and never return. That was a poor investment of our academic dollars. In the past, people did not leave because of the money; it was because they did not think Canadians were taking AI seriously enough. Five years ago, if an American venture capital firm were to invest in a Canadian company, it would ask that company to move to the US. Today, a venture capital firm will strategically invest in Canadian companies and ask them to stay in Canada. In fact, they are asking their portfolio companies in the US to move to Canada to establish AI research labs here. AI graduates are now finding that Canadians are taking AI seriously and that they can be at the world’s leading edge locally.
Some companies that have recently been funded by tier-1 foreign venture capital firms have received somewhere in the range of $40 to $100 million dollars. Prior to being acquired by TD, Layer 6 was still doing its fundraising round. When we were presenting our team in California, the reaction we got was, “Wow, this kind of a company could not have been built in Silicon Valley.” The talent shortage is so high there that all the talent would have been aggressively mopped up by one of the technology powerhouses. This gives us an opportunity to build a fresh economy with fresh capabilities from the ground up in Canada.