Canadian businesses have been slow to employ artificial intelligence, which could worsen the country’s long-standing productivity woes and leave it trailing competing economies, according to a new study from the Dais, a Toronto Metropolitan University think tank. 

Here’s what you need to know:

The output: Just 4.7 per cent of Canadian companies with 10 or more staff were using any kind of AI at the end of 2021. That put the country in the middle of the G7 pack, behind Germany, France and Italy; Canada outpaced the U.K., U.S. and Japan, although data from those countries is older. 

Low AI takeup continues “a trend that has been plaguing Canada for a long time,” said study author Angus Lockhart, a senior policy analyst at the Dais, in an interview. “We’ve never been good at adopting any kind of new digital technologies.” Canadian businesses spend relatively less on equipment, software and data than the OECD average.

Partly as a result, productivity has been flat or falling for decades. AI can help, Lockhart wrote, citing research showing it boosts firm- and industry-level output, although country-wide impacts are more mixed. (The Dais study doesn’t directly address automation-induced job losses, but calls for responsible use of the tech.)

Productivity gains create wealth, so AI adoption could “improve the quality of lives of Canadians,” Lockhart said. Plus, “we don’t want to fall behind and have our firms unable to compete.”

The reasoning: Most firms that hadn’t deployed AI had no use for it. Some 69 per cent of such companies said they hadn’t identified a business need for the technology, while 28 per cent were unaware of available tools. 

“The trouble is that businesses haven’t even tried to figure out what they could use AI for,” said Lockhart. But potential applications abound, he claimed, citing sector-agnostic needs like HR tools for time management and scheduling.

The report recommends governments introduce awareness programming. Lockhart suggested policymakers pay particular attention to small firms, and major economic sectors like manufacturing, which have been slower to adopt AI and risk falling behind.

The data: The study draws on Statistics Canada’s biannual Survey of Digital Technology and Internet Use, which polled 15,683 businesses between November 2021 and March 2022. It’s a robust data set—the measures of AI adoption that tend to get quoted have come from tech vendors, and draw from much smaller sample sizes.

Still, the agency’s numbers are two years old, predating the launch of viral generative systems like ChatGPT. “The next time we have good data on this, we’re going to see a pretty dramatic increase in the number of firms that say they’re using AI,” Lockhart said, predicting that much of that growth will be from generative tools used for office tasks.

“But that’s not the only kind of AI out there, and it’s certainly not the only kind that we want businesses to be adopting,” he said. Firms need to also take up tools like computer vision or complex machine learning systems that may require specialist staff to implement, but are proven to produce economic value.

Correction: This article has been updated to correctly reflect Angus Lockhart’s title at the Dais.