Artificial intelligence is expected to boost productivity over time, but its near-term impact on inflation may be more complex than policymakers anticipate.
BNN Bloomberg spoke with David Doyle, head of economics at Macquarie Group, about how AI-driven investment, labour trends and recent inflation data are shaping expectations for interest rates and economic growth.
Key Takeaways
- AI-driven productivity gains are expected to be disinflationary over time, but current effects appear to be contributing to inflation pressures.
- A surge in tech investment, including semiconductors and data centres, is driving higher costs across related sectors.
- Productivity growth has improved modestly since 2019, though gains remain concentrated in the technology sector.
- Labour market dynamics, including rising AI adoption and demographic shifts, are supporting gradual productivity improvements.
- Capacity constraints in skilled trades and energy demand from data centres are adding to short-term inflation pressures.
Read the full transcript below:
ANDREW: Kevin Warsh, the U.S. president’s pick to lead the Federal Reserve, has indicated he believes an AI-driven productivity surge would be disinflationary and justify lower interest rates. Canada has just released labour data, and the U.S. delivered its latest inflation report today, with inflation topping three per cent, the highest in two years, driven by energy.
Warsh, Mr. Trump’s pick for the Fed, has this “neutral rate” view. We’re joined by David Doyle, head of economics at Macquarie Group. Thanks very much for your time, David. Talk to us about what economists think will be the impact of AI. Will it bring down labour costs? Will it be disinflationary?
DAVID: I think where the view comes from on AI is that it will boost productivity growth over time. Productivity growth tends to be disinflationary and, as a consequence, there is a view in some quarters that the proliferation of AI and higher productivity growth will lead to lower inflation over time.
I would say our view is a little more tepid. As Chair Powell indicated in his March press conference, before you get to that disinflationary world, we’re still undergoing a period with a significant tech capital expenditure boom. There’s evidence suggesting inflationary pressures are building in tech. Even in today’s CPI report, for example, the category of computer software and accessories was up four per cent month over month. We’ve seen tech-related inflation picking up more broadly.
So while we can debate the extent to which AI will increase productivity and its long-term impact on inflation, where we are now is an interim phase where it appears to be contributing more to inflation than reducing it.
ANDREW: Yes, certain types of chips have become more expensive if you’re buying a PC because of AI demand.
DAVID: That’s correct. You can see it in the CPI report, but also in charts of the Producer Price Index for semiconductors, which had been in a prolonged downtrend — not just disinflation but deflation — for several decades. In the last year, it has turned sharply upward.
That’s an indication of the capacity pressures we’re seeing in the tech space, given the boom over the past couple of years.
ANDREW: And if you’re trying to build a non-AI factory in the U.S. right now and need an electrical contractor, you’re probably paying a premium because of all this buildout.
DAVID: That’s right. Data centre construction is another area that’s booming. You can see pressures in skilled trades. More broadly, U.S. retail electricity prices are up about five or six per cent year over year, which is a strong pace. The proliferation of data centres is a contributing factor and, if the buildout continues, it will likely add to those pressures.
ANDREW: Talk to us about U.S. labour productivity. What patterns are we seeing?
DAVID: Our team looked at this over the past few months. For a long time, productivity growth was subdued, particularly between 2010 and 2019. We think there is evidence that productivity growth has seen a modest pickup.
In the last cycle, productivity growth averaged around one and a half per cent. Since 2019, it has been a little above two per cent. There is a case that productivity growth is improving, though it may be concentrated in the tech sector for now.
If AI proves as beneficial as some suggest, that improvement could broaden across the economy. We are encouraged by developments in labour productivity, but there is still more progress to be made.
ANDREW: What about the number of adults using AI at work? Is that increasing significantly?
DAVID: It is trending higher. Usage is increasing, though perhaps not as quickly as some proponents might expect. We’re also seeing greater adoption at the executive level. It varies by occupation, but it is still early days.
We’re only a few years into AI becoming widely used, so we remain optimistic about where things could go. Another important factor for productivity is demographics. The Bank of Canada noted last year that demographics were a significant headwind to productivity growth from 2010 to 2019, but are now a modest positive.
That doesn’t mean demographics are a major tailwind, but they are no longer the drag they once were. A key factor is that the share of core working-age individuals — those aged 35 to 54 — is now rising after a period of decline, which supports productivity.
ANDREW: And what about older workers, those aged 55 and up, who are staying in the workforce? What does that mean?
DAVID: Older workers can be beneficial for productivity. They bring years of experience. There’s a perception they may be less technologically adept, but their experience is valuable.
Also, the share of older workers in the labour force was rising significantly in the last cycle and now appears to be stabilizing. That shift is also helpful.
ANDREW: We’ll leave it there. I hope you come back, David. It’s a fascinating area. David Doyle, head of economics at Macquarie Group.
---
This BNN Bloomberg summary and transcript of the April 10, 2026 interview with David Doyle are published with the assistance of AI. Original research, interview questions and added context was created by BNN Bloomberg journalists. An editor also reviewed this material before it was published to ensure its accuracy and adherence with BNN Bloomberg editorial policies and standards.
