Good morning. Is artificial intelligence good for the economy? In focus today, we look at how the tech’s emergence compares to earlier eras of disruption.

Energy: The B.C. government is backing a proposal to move more Alberta crude oil to the West Coast involving the Trans Mountain pipeline system, in a bid to stave off Alberta Premier Danielle Smith’s ambition for a new pipeline.

Manufacturing: Canadian fertilizer and agricultural giant Nutrien Ltd. has chosen a U.S. location for a project that could reach $1-billion – a blow to the Carney government, which has promised to attract sizeable investments in the mining and resources sector.

Investing: Questrade Financial Group Inc., one of Canada’s longest-standing challengers to the Big Six banks, is launching new trading products tailored to a growing wave of sophisticated retail traders.

Runaway hype around artificial intelligence, rising investor unease and unclear productivity effects have opened a divide among economists over what the new generation of AI systems actually means for the labour market and long-term growth.

The competing views turn on whether the latest wave of AI behaves like earlier technological breakthroughs, with productivity gains arriving only after costly reorganization, or whether its acceleration marks a shift in which automation outpaces the creation of new economic activity.

AI-related capital spending contributed 1.1 per cent to U.S. economic growth in the first half of 2025, outpacing household consumption in an economy typically driven by consumer spending, a recent report from JPMorgan said. The report noted, however, that data centres employ few workers once built – “especially compared to a factory or office campus” – limiting the job creation and local spending that usually follow large investments, a dynamic economists refer to as the multiplier effect.

The debate over whether this boom can generate the kind of broad-based employment and spending associated with earlier technological cycles is central to how economists are navigating early indicators.

‘History suggests this lag is normal’

Serdar Ozkan, a senior economic policy adviser at the Federal Reserve Bank of St. Louis, said AI investment follows a much longer pattern of general-purpose technologies that reshaped economies only after years of adaptation. From electrification to the rise of computers and the early internet, he argues, the visible promise of each wave arrived well before the productivity data showed any measurable effect, largely because firms needed time to redesign workflows and workers had to build complementary skills.

“History suggests this lag is normal,” Ozkan said. “General-purpose technologies require extensive time for experimentation, process redesign and work-force adaptation.”

“At the same time, the workers best equipped to adopt these tools may face the most direct competition from them, which makes this transition harder than the cycles we saw with computers or the early internet.”

‘This is pure job replacement’

David Rosenberg, founder of the Toronto economic consulting firm Rosenberg Research, said there is a key distinction between technologies that expand an economy’s productive capacity and those that speed up tasks that companies already perform.

AI is firmly in the latter category, he said – a shift he argues will compress labour demand rather than create new forms of economic activity. “This is pure job replacement,” Rosenberg said in an interview. “There is no offsetting multiplier impact.”

The types of work most exposed to rapid substitution include research, analysis and other information-processing roles, he said. Those tasks are increasingly performed at a speed that limits the emergence of new job categories around them, and they reach higher into the white-collar work force than entry-level positions.

Earlier technological cycles eventually increased output as firms reorganized and new industries formed, he said. But AI is narrowing that process by directly completing white-collar tasks, rather than enabling workers to do more.

Pain, gain, or both?

The uncertain benefits that AI might bring to business are reflected in the uneven returns it is paying. A new KPMG survey of 753 Canadian corporate leaders found that 93 per cent say their organizations are using AI in some form – up from 61 per cent a year ago – but only 2 per cent report a measurable return on those investments.

Compared to the United States, Canada’s economy leans more heavily on goods production and resource industries, which are less exposed to the white-collar substitution enabled by AI. “If you actually produce things that are tangible,” Mr. Rosenberg said, “you’ll be okay.”

The financial services sector here, however, faces pressures similar to those in the U.S.

“The typical Bay Street equity analyst isn’t producing stuff. They’re producing research for his or her boss, who now can get it for free and within a matter of minutes.”