Canada needs an industrial strategy for AI that prioritizes IP ownership, ICTC report says

Canada’s existing AI strategies have provided the groundwork to create a full industrial strategy to commercialize AI across the country, according to a new report from the Information and Communications Technology Council.

The report, released on Sept. 28, outlines three priority areas for an industrial strategy that would build on the current Pan-Canadian AI Strategy: IP ownership and commercialization, developing AI talent and the need for responsible AI frameworks. 

“Such a strategy can better catalyze technology transfer, grow domestic capacity for late-stage, IP-rich R&D, and promote Canadian IP ownership, while driving responsible AI innovation and generating sustainable, inclusive growth,” the report says.

The federal government introduced the Pan-Canadian Artificial Intelligence Strategy in 2017 with a $125 million investment, and renewed it again in the 2021 federal budget with a commitment of $443.8 million over 10 years starting in 2021-22. The strategy is being developed and led by CIFAR, a global research organization formerly known as the Canadian Institute for Advanced Research.

The Budget 2021 funding included $185 million over five years "to support the commercialization of artificial intelligence innovations and research in Canada."

The ICTC report notes that the strategy introduced in 2017 focused largely on academic research, but the 2021 budget commitments have expanded its scope and made it an “optimal time” to focus on commercialization.

Alexandra Cutean, the ICTC’s Chief Research Officer and one of the authors of the report, said that Canada is already a global leader in AI research, which provides a base to build from.  

“That's a great foundation from which to start. But the next step is how do we turn that into something that can produce economic value for Canada,” she said. 

Cutean and the other authors of the ICTC report found that acquisition was a common form of "exit" for Canadian startups, with the World Intellectual Property Office reporting in 2019 that Canada had the third highest number of acquired AI companies (ahead of countries like Germany, Israel, France and India).

The ICTC researchers looked at a sample of 209 Canadian AI startups that have exited, and found that approximately 50 percent were either acquired or bought out by another company. Of those companies that were acquired, 60 percent were bought by a foreign entity.

"How do we support companies to get to that later stage? Part of it has to do with additional funding, another part has to do with looking at the mechanisms that we have in place to support IP development," Cutean said, adding that there isn't a broad strategy across Canada for spinning off IP from research done at universities and post-secondary institutions.

Report calls for new indicators for innovation success 

In particular, the report emphasizes the need for new indicators to measure innovation success, both those related to IP rights and an inclusive digital economy.

The authors of the report noted that despite the importance of IP retention and commercialization, a recent assessment of the Pan-Canadian AI Strategy viewed foreign acquisitions as a positive indicator of success — despite the loss of IP to another country — and programs such as SIF and the Superclusters Initiative focus narrowly on jobs and GDP growth.

Cutean also said a robust strategy could use inclusive indicators to measure the potential harms of AI to society and implement policies to offset those harms.

"How are these investments promoting social good and what types of jobs are being created? What types of jobs are being displaced as a result? And do we also have policies or programs in place to effectively transition the folks that are being displaced into something else?" Cutean said.

"It's a little bit more robust rather than just saying, 'how much has our GDP grown and how many net new jobs are there?'"

R$