How does AI factor into growing Canada’s economy?

Artificial intelligence (AI) is increasingly seen as one of the most transformative technologies of our era. However, little is known about how widely it is adopted by firms, what sorts of firms are adopting it first or most effectively, nor how they plan to apply it. This lack of reliable information on the use of AI has made it difficult to formulate evidence-based predictions. This has not stopped people from wondering, and often worrying, about the role of AI in the economy and the future of work. My University of Toronto colleagues, professors Avi Goldfarb, Joshua Gans, and Ajay Agrawal, convened an influential meeting of the National Bureau of Economic Research (NBER) specifically on AI in Toronto in the fall of 2018. Some of the top minds in economics and management were there. We discussed and agreed on a lot of things, but mostly we left with more questions than answers. My main contribution was to raise the alarm about the need to measure AI and how important it is to do this at the very earliest stage of its diffusion. To that end, I have worked with professor Erik Brynjolfsson of Stanford University and researchers with the U.S. Census Bureau to examine this in the U.S. context. Some of our results apply to Canada and elsewhere. First, we have to recognize that AI is a very flexible technology. In fact, it is actually a class of technologies that more practically are discussed with terms like machine learning, natural language processing, voice recognition, and other “down-to-earth” applications. Yet, even these more-distinct applications are still quite broad, and they share a quality that academics think of as being a “general-purpose technology” (GPT). In a nutshell, this means that it is flexible, likely to diffuse quite broadly, and will have potential for high economic impact. Examples of GPTs in the past include the steam engine, electrification, and my personal favourite, the commercial internet. Yet, the challenge with these flexible technologies is that the firms intending to use them have to do a lot of inventing around them to make them productive. They have to specify how the technology should be applied, and how intensively, and what it should replace or augment. This, in turn requires imagination, innovation, and good old-fashioned execution. Even more daunting, firms often have to reshape the business activities and production processes that will leverage these technologies. It boils down to “it will take investment, time, and co-ordination.” In short, there is a gap between when the technology diffuses and when it shows up in the economy. Where are we on this journey with respect to AI? We know that there has been a lot of investment in AI-related technologies in recent years. Going back a few years, this investment actually starts with a broader move towards digitizing information, managing data, investing in infrastructure such as the cloud, and everything that has led to some firms even being able to deploy AI in a meaningful way. And even with this lead-up, the “hype” far outstrips the actual deployment, in practice. Kristina McElheran is an assistant professor of strategic management at the University of Toronto. Photograph courtesy of the University of Toronto How do I know this? This has started to be measured in the U.S. A new nationally representative survey, the Annual Business Survey (ABS), addresses this data gap. Brynjolfsson and I collaborated with the Census Bureau on the AI-related elements of that survey. The ABS collects information on the adoption and use of several advanced technologies—including those most closely associated with advances in AI, such as machine learning, machine vision, natural language processing, voice recognition, and autonomous vehicles—from a representative sample of more than 850,000 firms comprising the private non-farm economy. Our research found that average AI diffusion across the economy has been low. We estimate that the aggregate adoption rate for all firms in the U.S. economy was only 5.8 per cent (and 6.6 per cent when adoption is imputed) as of 2017, despite a lot of apparent investment. We are working on a new study with more details specifically in the manufacturing sector, which is a leading adopter of advanced technologies. How does this apply to Canada? The research finds the adoption is very concentrated in larger firms, and the distribution of large firms in the U.S. is greater than those in Canada. If there are any economies of scale in AI adoption, this will slow its diffusion north of the border. The other thing that is quite consistently emerging from my research on digital technologies, more generally, is that it is rarely enough to just think about the technology in isolation. Having other inputs that work with and reinforce the technology—what we call “complements” in the academic studies—are critical. For instance, I recently published a paper with Brynjolfsson and Wang Jin of MIT’s Initiative on the Digital Economy, showing that the use of predictive analytics spread very quickly in the U.S. But it only yielded economic benefits in firms that had other inputs in place such as a robust IT infrastructure, educated managers, and a production process that was relatively stable. Canadian firms should be looking at these findings and thinking about what other inputs will be important to make AI adoption achievable and productive. While we still need to continue the research, my intuition is that building the Canadian talent pool—not just in terms of developers but also managers and business leaders who can understand these issues of co-invention and co-ordination—will be essential to leveraging this fast-rising technology for growth in the Canadian context. Kristina McElheran is an assistant professor of strategic management at the University of Toronto. She is a visiting professor at MIT and Digital Fellow at the MIT Initiative on the Digital Economy. The Hill Times

How does AI factor into growing Canada’s economy?

Building the Canadian talent pool will be essential to leveraging this fast-rising technology for growth in the Canadian context.

Sudan’s warring sides and the long-simmering ethnic conflict

Balanced approach needed on right-to-repair bill, protecting IP rights and public safety

Reflections from a new citizen

Public Safety Canada’s dealing with natural and communication infrastructure disasters

As COP29 wraps up, will Canada finally stop funding fossils this fall?

Privacy and AI bill needs ‘right balance,’ and to pass in time for next federal election, says Liberal MP Ryan Turnbull

Canada has a strong head start in these early innings of AI

Canada needs to ramp up skills for AI revolution

Liberals fail to provide leadership on green and digitalization transitions, opposition parties missing in action

Canada’s innovation advantage

News

Get free news updates

Opinion

Submit a guest opinion article

resources

Advertise

subscribe

get help

Get free news updates

Submit a guest opinion article

How does AI factor into growing Canada’s economy?

Building the Canadian talent pool will be essential to leveraging this fast-rising technology for growth in the Canadian context.

Get The Policy Briefings Newsletter