Ep 44: The AI Race—Trillions In, Productivity Loading…

AI is supposed to make life cheaper, faster, and more productive, so why does inflation still feel stubborn and why do rate cuts still feel uncertain? 

Welcome to Part 2 of my conversation with Senior Economist at Indeed, Brendon Bernard. 

In this episode, we dig into a counterintuitive idea: the AI boom can be inflationary before it becomes disinflationary. When trillions flow into data centers, chips, and the electricity needed to power them, that investment can lift aggregate demand and add real cost pressure in energy and hardware. The payoff comes later, when AI actually reshapes workflows and pushes down the price of services people buy every day.

We also pull lessons from earlier technology waves, especially the long ramp from “computers are everywhere” to the late-1990s productivity surge. The benefits didn’t arrive on day one, and AI may follow a similar timeline. We talk about what it would look like for AI to finally lower the cost of legal work, accounting, financial advice, education support, and other knowledge-work services, and why the “human touch” still matters in roles built on trust, coaching, care, and interpersonal judgment.

Then we zoom in on the labor market. For many experienced workers, things can look steady even in AI-exposed occupations. For younger workers trying to get a first foothold, the picture is more complicated: weaker postings in highly exposed roles, a post-pandemic tech hangover, and a tough question about causality as interest rate hikes overlap with the rise of generative AI. 

We close with the policy challenge: how do governments plan for re-skilling and economic security when we don’t yet know whether AI mostly augments workers or automates whole task bundles, and whether today’s safety net is built for long-term income risk?

Subscribe, share, and leave a review, and tell us what you’re seeing firsthand: is AI helping your work, changing your job search, or shrinking entry-level opportunities?

Quick heads up. This episode was recorded on February 19, 2026 so while the news may have changed since this conversation was recorded. The thoughts and ideas still remain relevant. 

Everything in this episode is for educational purposes only. Not financial advice. 

Transcript

Why Inflation Still Feels Sticky

Julia Pennella, Host 0:00

AI is supposed to make everything cheaper, faster, and more productive. So why does inflation still feel kinda sticky? Welcome back to Let's Talk Politics, the podcast where we break down the economics and policy shaping your daily life and help make politics make sense. This is part two of my conversation with senior economist at Indeed, Brendan Bernard. In this episode, we continue the conversation on the AI boom. Trillions of dollars are being poured into the AI race, from data centers and chips to the infrastructure needed to power AI's massive electricity demand. And all that investment might actually be holding inflation higher before the productivity payoff can start showing returns. We dig into what history can teach us from the late 1990s IT boom to today's AI wave and why the benefits of new technology can often lag years behind the hype. And then there's the labor market. For experienced workers, things still look relatively stable. But for younger workers trying to get their foot in the door, the outlook is getting way more complicated. So the big question on everyone's mind: will AI help workers or replace them entirely? Because that answer will shape everything from wages and reskilling to how governments rethink economic security. Quick heads up. This episode was recorded on February 19, 2026. So while the news may have changed since this conversation was recorded, the thoughts and ideas still remain relevant. Also, everything you hear in this episode is for educational purposes only, not financial advice. So, let's stop politics and the economics shaping them. A bit of a technical question here. So for any of the uh economists and really number crunchers listening, I want to talk about the impact AI has with setting the central bank rates. Central banks typically view technological progress as disinflationary. However, what we're seeing with AI and that data center investment is now accounting for roughly 1% of the US GDP and driving up energy and chip costs. So are we in phase one where AI is actually a pro-inflationary force? Or what else uh do you think is at play here?

Which Services Get Cheaper First

Brendon Bernard, Senior Economist at Indeed 2:27

Yeah. And so this is a this is a good uh question. I this would be a good question to put on like a uh undergrad macroeconomics exam. So if there are any uh economic professors on online, I I think this would be a good one. This is like an empirical question, and I'm sure there are uh lots of studies that are coming out on this. I think you put it well in terms of like what phase of the boom we're in. One of the historical experiences provide some support for the disinflationary effects of technology is what the US economic boom in the late 1950s, when the US unemployment rate got quite low and the US economy was really chugging along, and yet uh inflation was relatively subdued. And the Fed didn't hike necessarily as fast as uh it would have otherwise, uh, in part because overall it looked like the inflation environment was manageable, even as the economy was in an economic boom. Thinking though, of like where things were in that economic cycle and how that compares to today, the IT revolution and computers in the economy that didn't just start in the late 1990s. That's probably though when we saw started to see the payoff in terms of goods and services across the economy, having cost pressures lowered by this new technology. At the same time, though, there that requires a massive build-out, a massive amount of investment uh to happen, that every office in the country basically needs to get a new computer system. And so that was a trend that I think predated the actual benefits and productivity boom that eventually came in the the late 1990s. And so, thinking of where we are today, it's quite possible we're still in that first phase, in the sense that there's clearly a ton of investment happening uh in this area of the economy. Has it fully though translated to the price of services, especially using these types of goods, to really come down? I'm thinking the price of legal services, the price of accounting services, the price of in advice and investment services, all these different things. It's probably too tuned to say that. Instead, we're in phase one, the build-out, where this is clearly a rise in demand for lots of things, including uh electricity. If it's boosting aggregate demand, then all else equal. The central bank needs to be aware of that and be thinking about if not like raising interest rates, at least be of the mind that this is strong demand, is inflationary in that uh in policing, at least needs to take that into account among the other uh various things. And we see in the US US, for instance, that uh inflation is still somewhat elevated. In Canada, it's come down a bit, but Canada doesn't have the same aggregate demand boom coming from uh rising investment. So I think in that sense, uh the the Fed and uh and other central banks, they of course are thinking about uh all this to expect that, oh, we're gonna all of a sudden have now have a deflationary force that uh actually necessitates cutting rates to keep things stable. I think we're um aways from that, but we could get there once we start seeing the types of goods and services that are really heavily using this technology start to get cheaper. For the average person, that's where the payoff from this economic growth is gonna come from. And it's a fact about automation and technology over time that we get better at making stuff, and often that means fewer people are needed to produce a good, starting from the very basics, which is food. We have way fewer people producing food than we used for us to, but the cost in terms of like an hour's work to purchase all the food you need magnitudes less than it used to be. So long term for the AI to benefit the average person, this is how it's gonna sh show up. All the services uh that they use that are involving computers and by extension now, AI are gonna get way cheaper. Their their purchasing power of their wages rises in response.

Julia Pennella, Host 6:39

It's is really interesting, the service sector to watch, at least the next year.

Brendon Bernard, Senior Economist at Indeed 6:44

You know, in that example of consultant, you can also lump education into that as well. But the thing is, though, that's kind of an example where these days information is so readily available, you can search whatever you want and often get good responses. That's different than having a coach or teacher actually work through with you what that takes. And so the human touch still still matters. And in some ways, those jobs remain very important, both consultant, but also teacher, but like educator, motivator, fitness coach, like all these things they're all um still important, but they they're gonna change uh because this technology and we add medicine to the question as well. I know this is a plot line on uh season two of The Pit, uh, which I had just started watching. The human touch is necessary and it's not gonna go away, especially in many fields where interpersonal contact is important, but the this new technology is gonna have a transformative impact in other ways.

Julia Pennella, Host 7:44

When we look at interest rates and economic growth, Canada and the US aren't always moving in sync. So, how are central bank decisions on interest rates showing up in GDP growth? And is the Bank of Canada being more cautious than the US Federal Reserve right now? And in addition to that, do you think the US could move faster to cut rates than the Bank of Canada is?

AI And Hiring Signals In Data

Brendon Bernard, Senior Economist at Indeed 8:10

The Canadian economy has been weaker than the US economy for a few years now. It shows up in the labor market data, so it's definitely shows up in GDP growth and also shows up in inflation, where Canadian inflation is much closer to target than the US. The US economy is going to impact Canada regardless. One thing that's interesting, if you can't compare Canada and the US, we have might have different central bank policy rates, but the 10-year government bond rate in Canada and the US fluctuates pretty closely. So US policy is gonna impact Canadian rates regardless of what the Bank of Canada does. And so whether the Fed cuts or not, it's all about what the implications of that for the Canadian economy.

Julia Pennella, Host 8:55

How have you seen the labor market evolve over the last few years as we've seen more exposure of AI in workplaces? And what does that impact have on our economy? What does this mean for those longer-term consequences?

COVID Digitization And The Tech Hangover

Brendon Bernard, Senior Economist at Indeed 9:09

So I think right now, fitting with a lot of the GDP data, that um it's tough to really pinpoint particularly large labor market effects of AI in Canada right now. Some of the changes that we've seen, they're tough to separate from what happened previously during the post-pandemic boom. So I I've been looking at occupations in the labor force survey data where there's high exposure to um AI, and specifically thinking about high-wage jobs. So professional and managerial jobs, jobs where computers are an important tool, jobs that are paying well, jobs that we are worried about long term because AI clearly can replace human workers in many of the tasks involved in those roles. What's going on in employment in those areas? Overall, if you draw a line where end of 2022, before and after, you don't see a major effect. And if and those jobs are also growing quite rapidly before 2022. There's a potential effect, but once again, tough to kind of uh tease out the trends, is particularly among younger workers. So in these high exposure jobs, I compared uh the share of the population employed in these roles, over 30 and people in their 20s. And what we see is among the people over 30, their their employment is steady, pretty elevated. The share of the population in these jobs is pretty stable over the past few years. Among those in their 20s, uh, it's declined a bit since the end of 2022. That is a potential effect. But there is important context, which is that the boom in these jobs prior to 2022 was particularly large among those in their 20s. And so young people really benefited from the Canadian tech boom and boom in other white-collar jobs as well during the years or so leading up to now that's reversed since. And it's reversed at a time when clearly we're in the new AI age. But how much is AI actually causing that? Like thinking about employer demand and hiring appetite. We see on Indeed in both Canada and the US that compared to where things were in 2020, job postings in highly exposed occupations are weaker compared to earlier 2020 than they are for other job types. So also raising red flags. The thing is, though, that a lot of this weakness really showed up in basically the six months before ChatGPT went public. Really, the turning point, actually, just timing-wise, coincidence or not, was when central banks, including the Fed and the Bank of Canada, started hiking interest rates. And so that sort of like highlights the challenge of teasing out these effects. But I think in aggregate, given the muted impact that AI has had on just like the Canadian GDP numbers, these limited employment effects probably shouldn't be surprising. But that that doesn't mean to say that some of the weakness in hiring among youth isn't related to the rise in AI. And definitely doesn't say anything about what to expect going forward, because I think in a lot of ways, we're just like at the cusp of how AI is going to transform how employers are structured and the types of tasks they delegate to workers. And so maybe we're just too soon uh to really start seeing effects.

Julia Pennella, Host 12:40

One last question for you. It was really interesting how you said the timing of when Chat QB went public and we were seeing interest rates rise. Well, we won't know the actual answer, but I'm just curious, like hypothetically, do you think COVID slowed down the potential AI boom in the impact of the labor market? Or did COVID prepare us for potential of what labor loss is going to look like in the next maybe 10 or 20 years because we had that mass layoff between, you know, lower productivity, less people working at the time because of COVID? Like I'm curious how COVID fit into that equation, if it's either a positive or a negative. Again, just from your observation.

Brendon Bernard, Senior Economist at Indeed 13:18

The COVID job losses, I don't think really track too much to the current prospect of automation going forward. Um, in part just because literally on net, the job losses happened in two months, March and April 2020. And then it was just kind of like a rise back from there. And it was all related to the pandemic and the restrictions in response. And once things returned, jobs returned quickly as well. One of the real outcomes from the initial pandemic boom, though, was that the Canadian tech sector and other Canadian white collar jobs, but particularly in tech, they were rising pretty quickly pre-pandemic and then went into overdrive in years after. There was a move to digitize even further. The boom in work from home, boom in e-commerce, all these things. The economy further digitized. I think that digitization accelerated during the pandemic. And that acceleration was coupled with a boom in hiring. So we did digitized, but employment also went into that sector. Now, many of those jobs are some of the most vulnerable to potential automation effects, both because the tasks and skills involved in those jobs have a lot of overlap with AI. Anyone who works at a computer has at least some tasks and skills that are uh related to the capabilities of AI. But also because, at least in the tech sector, that boom in employment was concentrated among employers who are probably gonna be the most likely to adopt the technology first. They're more familiar, more comfortable with new technology and AI. And so, in that sense, there was even absent AI, there was probably gonna be a hangover in hiring after the kind of post-pandemic boom ended. And that was probably the thing that kicked off that real sharp drop in tech and other white-collar job postings in mid-2022. Now, though, it leaves us in a bit of a precarious position where the share of the population working in jobs that are vulnerable is actually greater than it used to be because of that digitization went into overdrive. And now I think employers like haven't cut back so much. You know, as I mentioned, those over 30 are uh their their employment in these jobs is doing um quite well. Uh, but you know, if they stop hiring, well, then for job seekers, it's tough to land those new roles. And though this is probably uh a factor that's making it difficult for young people and other uh folks looking to move up the career ladder to land those jobs. And this is where we kind of stand today. Um, and we haven't even really seen many of the major effects uh from the technology that could easily come in the next few years.

Julia Pennella, Host 16:14

What's the biggest mistake governments are making right now in preparing workers for AI? Are they not investing in reskilling? Are they underinvesting in it? Should we be looking at income supports? What would what do you want to see the government do as we're moving into this AI stage and how the job market's gonna fluctuate?

Brendon Bernard, Senior Economist at Indeed 16:30

I think when it comes to specifically responding to AI, I hate to say it, but I don't have a great answer. Uh in part because going back to something we were talking about earlier, how is AI going to be used three years from now? I don't know. And and we should have some humility around our predictions uh about that, because what the potential policy response is actually really depends on how AI is being used. Is it the sort of companion, the augmenter, the thing that uh helps workers uh in their job uh and and changes the world of work, but in a more subtle way than the kind of agentic world of automating many tasks that people do, potentially some full roles. I think like the the policy prescription there is different and the direction of technological change is unclear. That makes crafting the right policy even tougher. And so I I'm I'm not exactly sure of the answer. I think at the end of the day, though, that doesn't mean there aren't lots of economic policies that would be worthy of undertaking regardless of the AI boom. And as we see in the relatively flat GDP numbers over the past few years, there's a lot of basic things that we can work on. And even as we're potentially entering an era that is pretty transformative.

Julia Pennella, Host 17:54

Well, I like that. And you gave me a lot to think about as I kind of bounce this idea with my policy friends, because I'm wondering to your point if it's going to be policies reactive to how society and technology and the economy is adapting versus prescriptive and ahead of the curve. So interesting.

Brendon Bernard, Senior Economist at Indeed 18:09

Yeah, I one thing that I'm not an expert on this policy, but thinking about insurance and the government's role there, employment insurance in Canada, it's designed for covering certain risks that appear pretty different than in the world of AI, where, like I said, we have no idea the extent of how certain careers are going to be transformed. Future labor market income is potentially significantly at risk from new technology. What type of insurance policy do we have today to deal with that? We have basic social assistance, we have employment insurance if you get laid off. That's a different risk, though, than long-term income security. And these policies aren't necessarily easy to design, but I would love to listen to an episode with an expert on this topic who's thought a bit more about the policy implementation and design of income insurance over the long-term, something a little different than employment insurance as currently constructed.

Wrap Up And How To Support

Julia Pennella, Host 19:13

And I work with an organization called Moms at Work. And that's one thing we talk about a lot too, is like unemployment insurance for moms specifically. Like in Ontario, if you get laid off after maternity leave, you've used up and eaten up your EI. And your EI also doesn't sustain you with what mortgage prices are now and the cost of living. So really interesting. Yeah, I'll circle back with you and hopefully uh get a guest on that. I think the broader conversation is the government should be having these types of conversations or consulting with these experts. And from my perspective, it doesn't seem like they are there yet because I mean there's a lot of geopolitical bigger things at play. But um, all that to say, Brendan, it's always a pleasure to have you on. We're gonna have to have you on again uh as this keeps evolving. But I really appreciate your insight. Well, that's a wrap on this episode. Thanks for tuning in for my two-part series with senior economist at indeed Brendan Bernard. If you like this episode, make sure to hit that like button, subscribe, leave a review, or share the episode with a friend. It really helps support the podcast and keeps these conversations going. I'm your host, Julia Pinella. This is Let's Talk Politics, and we'll catch you on the next episode.