Bot the difference: AI’s absence in economic data

0:00Jason Palmer:That new thing? Yeah, we've got it. The Drop by GNC, bringing you all the newness that matters. Handpicked by the pros who actually know what's up and what's proven to work. We keep you on top of the trends and dialed into what's next. Whether you're crushing it at the gym, leveling up your game, or thriving every day, The Drop by GNC is where the latest solutions in health and wellness land first. Non-stop innovation and fresh finds daily. Explore what's new and what's next on The Drop by GNC. Exhausted from spending half your night cooking, Factor eliminates the stress with chef-crafted, fully prepared meals delivered to your door.

0:38Jason Palmer:Fresh, never frozen, ready in two minutes. Factor delivers zero cooking, zero stress. Just heat and eat over 100 dietician-approved options weekly with no refined sugars, artificial sweeteners, or refined seed oils. Right now, go to factormeals.com slash easy50off and use code easy50off for 50 % off and free breakfast for a year. That's factormeals.com slash easy50off and use code easy50off.

1:15Alex Domash:The Economist.

1:22Alex Domash:Hello and welcome to The Intelligence from The Economist. I'm your host, Jason Palmer. Every weekday, we provide a fresh perspective on the events shaping your world.

1:36Alex Domash:A horrific attack in western Nigeria earlier this month is just one sign of a troubling change. Jihadist groups are splitting, leading to more violence between them, and spreading ever closer to the country's urban centers, threatening violence for all.

1:55Alex Domash:And nobody ever gave Virginia Oliver any hassle for being a woman running a lobster boat. No one dared. We look back on a career spent working Maine's waters for nearly a century.

2:14Alex Domash:But first...

2:24Alex Domash:Your favorite economist and mine, John Maynard Keynes, made one wild assertion back in 1930 in his essay Economic Possibilities for Our Grandchildren. He does a potted history of humanity, pointing out the incredible pace closer to his time of what he keeps calling technical inventions and technical improvements. They'd had huge impacts on workers' productivity. He concludes that by 2030, we'd all be working 15-hour weeks. I don't know about you, but four years out and that still looks unlikely. People love pointing out this folly of the great man, but let's take a broader lesson. Big technical improvements, like, say, artificial intelligence, take maybe a little longer than you might think to have big economic outcomes.

3:14Jason Palmer:AI capabilities are certainly improving very fast, but the effect of AI on the economy, not so much.

3:23Alex Domash:Alex Domash is our economics correspondent.

3:26Jason Palmer:AI may well lead to a productivity boom one day, but that productivity boom is not here yet.

3:34Alex Domash:And you say that because you've been digging into the economic data. What are they saying?

3:39Jason Palmer:Looking at America, there has been a sort of puzzle in the macroeconomic data, especially over the last year. So throughout most of 2025, you had on the one hand a booming economy. You had real GDP growing quite rapidly for most of the year. And at the same time, you had a slowdown in hiring. You had pretty sluggish employment growth. And especially in the second and third quarters in America, real GDP was growing quite rapidly. In the fourth quarter, we did get a real GDP print that was lower than expected. It came in at 1.4%, which sort of tempered the narrative of this big GDP boost with slow employment.

4:17Jason Palmer:But still, throughout 2025, you did have real GDP growing at 2.2%, while you had employment growing at an average of 15 ,000 jobs per month, which came out to about 0.1 % employment growth through the year. So the fact that there was this big gap between real GDP growth and slow employment growth, it usually would imply that productivity growth is quite high and that workers are producing more with less hours worked.

4:42Alex Domash:You say the gap between those numbers would normally be attributed to a growth in productivity per worker in a way that suggests that's not the explanation here.

4:50Jason Palmer:Correct. So 2025, of course, was a unusual year for many reasons. But looking first at the output side, of course, we had these massive expenditures in artificial intelligence. Firms were investing in AI infrastructure. And studies have shown that this did largely boost real GDP growth in America. At the same time, on the employment side, there were some funny things that were going on there as well. You had tighter immigration policy with the Trump administration. You also had this phenomenon with a lot of temporary workers that were falling out of the labor market. And what this did was basically, one, it's artificially constricted employment growth, but it also meant that the types of workers that were getting pulled out of the labor market tended to work in lower productivity sectors.

5:38Jason Palmer:I will also say that while a gap between output growth and employment growth sounds unusual, actually, when I looked at the data since 1950, in one-third of the years, the gap between those two was actually two percentage points. So by that metric, what we actually saw last year was not all that unusual.

5:58Alex Domash:Okay, so if instead we are looking for solid measures of the contribution of AI to productivity, where should we be looking?

6:07Jason Palmer:It's not a straightforward question, but one intuitive estimate that you could do is you could, one, look at what does research say about adoption of AI among workers? Two, among the workers that are using AI, how frequently are they using it? Are they using it every day? Are they using it sparingly? And the third thing that you'll want to look at is what the research says about the productivity gains from actually using this technology. Let's start with the adoption that you mentioned. We already know that's fairly high. Yeah, so the adoption is high and it's been picking up over the past year.

6:39Jason Palmer:And it is sort of striking that across a number of different studies, they all find almost the same thing. And that's that four out of 10 working age Americans are using AI on the job. So we're pretty comfortable that in our estimate for AI adoption is around 40%.

6:54Alex Domash:Okay, and then it was the intensity. People are using it, but how much?

6:57Jason Palmer:So the intensity is a little bit surprising. They're not using it as intensively as you might think. So only 13 % of Americans are using AI every single day on the job. And some recent research tried to quantify total hours where workers are using AI in the workplace. And this typically comes at about an average of two hours per week across all workers, which is about something like 5 % to 6 % of total working hours workers are using artificial intelligence.

7:27Alex Domash:And then what seems to me to be the real nub of the question, when people use it, how much it gains them, basically.

7:35Jason Palmer:Yeah, and so here there's been a number of studies over the past couple of years that have tried to answer this question. Some have looked at BCG consultants and gave them a lot of real-world exercises and tried to measure, are they doing the tasks quicker using AI? Are they able to produce their PowerPoints, produce their writing, produce their business plans more efficiently? And other studies look at in a legal context, others look in a writing context. And a synthesis of this literature found that on average, workers using AI have seen efficiency gains between 15 to 30 percent.

8:07Alex Domash:Okay, so taking it all together then, I mean, those all sound like fairly positive numbers. These are all pointing towards productivity gains. When does the massive boom that people are talking about actually start?

8:19Jason Palmer:Yeah, so if you do a back-of-the-envelope calculation using these three numbers, so 40 % adoption at an average of two hours per week, at an average of 15 % to 30 % efficiency gains from using the technology, what you find is that productivity over the past year would have increased between 0.25 and 0.5 percentage points, which sounds very small, but it's actually not insignificant. However, that's almost certainly an overestimation, One, that assumes that every time a worker is using AI, they are using it to its full productive capacity and using it to increase their efficiency. And two, it assumes that all of the time saved from using AI is then getting redeployed to other productive tasks.

9:04Jason Palmer:And I think anybody thinking about their own experiences in the workplace knows that this is hardly a realistic. If something makes my job easier and take less time, I'm going to knock off early. And fascinatingly, there was actually a study that came out over the past couple months that looked at tech workers using AI. And maybe this isn't that surprising if you've been following the news, but tech workers are working more than ever. So AI might be increasing their productivity and those increases in productivity have actually gone into working more hours, experimenting more with the newest models, with cloud code and vibe coding and all of that stuff.

9:39Alex Domash:So what's the take home message here? Is it the issue that adoption is still on the rise and people are still figuring out how to use it and that we're too early in the story to say what the ultimate productivity tale is?

9:50Jason Palmer:So the real message is that when you look at history and you look at when an economy actually sees a productivity boom, the lesson is quite clear that productivity gains actually occur when firms reorganize their production around the technology and when they start adopting new business models rather than workers just using the technology more. For For example, when electricity first arrived, you didn't see a big productivity boost from simply replacing steam engines with electric motors. Where you actually saw the productivity gains was when floor plans in factories were redesigned to capitalize on electric power.

10:29Jason Palmer:The same occurred during the computer age as well. Of course, many will be familiar with Robert Solow, a Nobel Prize winning economist who had a quote that basically the computer age could be seen everywhere except in the productivity statistics. And the reason was that simply using computers was not what was actually going to lead to the big efficiency gains. It was when companies actually restructured and reorganized and retooled their operations around this technology. And so organizing businesses around artificial intelligence is when the real productivity gains will occur.

11:02Alex Domash:Alex, thanks very much for joining us.

11:04Jason Palmer:Thank you for having me, Jason.

11:15Jason Palmer:Youth mental health is a complex challenge that requires comprehensive solutions. We must strengthen after-school programs. We must make digital literacy tools available in our schools. We must work with mental health professionals to support children. And we must empower mentors, educators, and parents to keep kids happy. Learn more about our commitment to finding lasting solutions at empowerourfuturecoalition.com slash solutions. Paid for by the Coalition to Empower Our Future. Most all-in-one HR systems are a patchwork of disconnected and manual tools. Rippling is totally automated. If you promote an employee, Rippling can automatically handle necessary updates, from payroll taxes and provisioning new app permissions to assigning required manager training.

12:01Jason Palmer:That's why Rippling is the number one rated human capital management suite on G2, TrustRadius, and Gartner. If you're ready to run the backbone of your business on one unified platform, head to rippling.com slash AcastBiz and sign up today. That's R-I-P-P-L-I-N-G dot com slash AcastBiz to sign up.

12:28Alex Domash:On February 2nd, as every evening, the residents of Kayama in western Nigeria followed the call to prayer. Many never made it to the mosque. Armed men brutally attacked two villages in the region. Residents were shot at close range or their throats slit. Some were burned alive.

12:48Ọrẹ Ogunbiyi:They set them ablaze. You can see for yourself, but we did them no wrong.

12:54Alex Domash:Nigeria's president has since sent troops to the region, including Brigadier General Nicholas Ruhm.

13:00Ọrẹ Ogunbiyi:Our aim is to stabilize the area and then expand our operations at work to ensure that we track and possibly strike the people who did this.

13:10Alex Domash:America announced it's sending military support too. But the bigger problem is not just the violence in Kayama. It's that it reached Kayama at all.

13:19Ọrẹ Ogunbiyi:For years, Nigeria has faced several interlocking security crises, especially in the northeast and the northwest.