¹ú²úÓÕ»ó¸£Àû

"A lot of the times when I start to deal with a Canadian company, they say, 'We've got to get an AI strategy. We've got to build AI,'" Frosst said at the University of Waterloo's Tech Horizons Executive Forum in Toronto on Tuesday.

"Then, it takes a long time to get from some high-level room that says we need this thing to an actual implementation that's sitting in production, saving their employees time or ... delighting their users."

Nicole Janssen, the co-founder of Edmonton-based AI firm AltaML, has had a similar experience.

She estimates it takes 18 months for companies reaching out to her business to commit to using AI and then another 18 months to start doing something with it.

"Then people get tired of this thing that's not giving them a return on the investment and it falls to the wayside," she said.

Tech leaders have long lamented the slow rate of adoption for their products in Canada, especially when compared to the U.S.

Some have blamed the pace on a lack of funding, while others have said it's a matter of culture.

Frosst said it's hard to narrow down what's hampering the rate of adoption.

Culture could be part of it, but he said, "I want to be clear that I don't necessarily think that cultural thing is bad."

"Some of the things that I really like about Canada is that we're slow and a little conscientious," he said.

"But it also has downsides and one of the downsides is five quarters of real GDP per capita decrease."

Those GDP declines have sparked a discussion about whether Canada is facing a crisis in productivity because it is lagging behind the U.S. and many other Nordic nations.

Frosst estimates large language models — the underpinning of AI, which use massive data sets to recognize, translate, predict or generate text and other content — could make a big dent in Canada's productivity woes.

He said LLMs alone will "augment" about 20 per cent of knowledge-based jobs, which include teachers, doctors, financial analysts and marketing consultants.

But to ensure LLMs and AI are "an absolutely massive opportunity" for Canada, Frosst said the country must not squander the foundations that have been laid for it.

Canada, for example, has long been known as a hotbed of AI innovation because of its focus on AI research and talent development.

Much of that work has happened through the Vector Institute and Mila, AI organizations based in Toronto and Montreal, respectively, which AI pioneers Geoffrey Hinton and Yoshua Bengio are deeply involved in.

Cohere has received funding from Hinton, who recently won a Nobel prize, and Frosst was one of his proteges.

"Some of the best minds are still here, some of the best institutes ... are here, but we have fallen behind in adoption," Frosst said.

At the same time, every other nation is gaining ground.

"It's kind of table stakes at this stage," he said.

"America is doing it, the whole world is figuring out how to increase productivity with large language models, and although that technology came from here, we've been a little delayed in adopting it."

To reverse the problem, Janssen urged business leaders to get moving — and quickly.

"Don't ask the question, 'What am I going do with AI?' but, 'What am I going to do with AI by the end of the year?'" she said.

"Because if we don't get started, we are going to fall behind and our productivity challenges are going to be so much more."

This report by ¹ú²úÓÕ»ó¸£Àû was first published Nov. 26, 2024.