Americans are using AI at fairly high rates. What does this mean for the economy?

Before conducting a recent survey to find out how much Americans are using generative AI, Harvard University economist David Deming said he was in the “AI skeptic” camp. That is, he was skeptical that the explosion of generative AI would bring major benefits to the US economy anytime soon. But now, he says, he is more optimistic.

“I was very surprised by the numbers in our survey,” says Deming. “And it kind of made me think that AI is going to be a bigger deal than I thought.”

The study, Deming said, was motivated by questions about whether and to what extent Americans are using generative artificial intelligence. Doing what economists tend to do, he and his colleagues Alexander Bick and Adam Blandin wanted to get some good data.

They modeled their survey after the Current Population Survey (CPS), which is sponsored by the US Census Bureau and the US Bureau of Labor Statistics. Conducted monthly, the CPS is like the gold standard for surveys. That's how we know things like the unemployment rate and the state of the labor market.

In short, the survey conducted by Deming, Bick, and Blandin is of high quality. It has a large sample size. It is a national representative. And they even included some of the questions that CPS asks, so they could cross-reference their survey with CPS and make sure their numbers were good. They conducted their survey twice in June and August 2024.

Deming said he was shocked by the results. He and his colleagues found that about 40% of Americans, ages 18 to 64, have used generative AI. And a large percentage seem to use it regularly. In their August survey, the Economists found that more than 24% of American workers “used it at least once a week prior to the survey, and one in nine used it every workday.”

Even more surprising, Deming says, the use of AI seems to be pretty high across the board. He expected younger and more educated Americans to be the biggest users. They confirmed it. “But we even found that 22% of blue-collar workers say they use AI, and usage rates were above 20% in every major occupation category except personal services, where it was as low as 15%,” he says.

This rapid rate of adoption, the authors noted, was much faster than the speed at which Americans adopted personal computers and the Internet. This might not be too surprising, though, considering that personal computers cost tens of thousands of dollars in the 1980s and 1990s, and you had to get the physical hardware, connect it, figure out how to use it, and so on. Even with the internet people had to do. Buy modems, get subscriptions, figure out how to “surf the web” and more. Generative AI is more plug and play, often free or has a low-monthly subscription cost, and has a user interface that's familiar to anyone who's used Google.

What does this mean for the economy?

In a recent two-part series indicatorDarian Woods and I debated whether AI is overrated. Given the tremendous uncertainty about the impact that AI will have on the economy — as well as our reluctance to make predictions that might laugh at us in a few years — we've decided to blurt out our personal feelings on the matter. We flipped an AI-generated coin. I got “AI Overrated” and wrote an episode and a newsletter that made a compelling case.

During my research, I discovered a survey conducted by the Census Bureau this year that found that only 5% of American businesses said they had used AI in the previous few weeks. I used this as evidence that the use of generative AI was pretty pathetic, especially considering all the hype surrounding it.

How does Deming square his findings with that finding?

For one, Deming said, he and his colleagues found that Americans actually reported using AI more in their personal time at home than at work.

hmm ok Putting my “AI overrated” hat back on, it makes me think that most of the uses of AI are not for productive work but for pleasure – which suggests that AI's impact on the economy will be limited.

The economic dream for AI is that it will rescue the US economy from its long period of sluggish productivity growth. Increased productivity — meaning workers are able to produce more in less time — is the magic sauce for raising living standards. And recent technologies have been pretty disappointing on that front.

I mean, look at smartphones. If I were to tell you back in 2006, the year before the iPhone was released, that soon we would all have supercomputers in our pockets, capable of searching the Internet, giving us directions to any location, sending emails, and even making video calls. -Employees and clients, basically order any product or service, translate the language and on-on-you'd think we'd see an explosion in productivity. But smartphones seem to be more of a tool for pleasure and distraction than an incredibly powerful work tool. We haven't seen a huge increase in productivity growth since its widespread adoption a decade ago.

Deming, however, emphasized that people are using generative AI for work. He says their survey shows that “about 1 in 4 people used it at work at least once in the past week” (vs. “1 in 3” who used it in their off time). When it came to how people use AI to help with their work, respondents indicated that writing, explaining, and administrative support were most helpful (although a significant portion also said it was for other tasks, including coding and support when working with customers and colleagues. using).

Squaring his study with one from the Census Bureau, Deming said he thinks “there's a big gap between the use of AI and a company's formal policy regarding employees. I think a lot of uses are under the radar, like when you have to write an email to your boss and You use it to write emails faster and you use ChatGPT to do it but you don't tell anyone.” Deming says that many of his students use generative AI in this way.

We reached out to Massachusetts Institute of Technology economist Daron Acemoglu — who, these days, you might call a major AI skeptic — about the research by Deming and his colleagues. Does the fact that a large percentage of Americans are using generative AI change their view at all?

“My concern with their numbers is that it doesn't distinguish fundamentally productive uses of generative AI from occasional/pointless uses,” Acemoglu said in an email. “If you're interested in what ChatGPT says to introduce one of your guests and use it to get a feel for it, you're a user, but it's not fundamentally integrated into the NPR production process. Even worse, if you ask me if I'm using generative AI, and if I'm honest, I'd have to say yes, because when I search I see generative AI-generated output but its productivity improvements won't require fundamental integration with complementary investments in organizational capabilities and workforce skills and Reengineering the production process – We know ChatGPT has about 200 million unique monthly users, but the question is how many of them are using it. This will lead to significant productivity improvements/cost reductions. I don't really know the answer to that question, and I don't think so. We will find the answer to that question in this paper.”

How much will generative AI increase productivity?

To try to estimate how much AI would increase productivity, Deming and his colleagues did some rough calculations. They looked at five randomized studies that analyzed whether the use of generative AI increased productivity in a variety of work settings. They picked the productivity-boost number in the middle of that study, which is about 25%. Then they multiply that, Deming says, by “our estimate of the number of work hours currently assisted by generative AI” (which, they calculate, is between 0.5% and 3.5% of all work hours in the US).

They concluded by estimating that generative AI would result in “labor productivity increases of between 0.125 and 0.875 percentage points at current levels of utilization.” That may not sound too good. But, Deming says, consider that productivity growth over the past several decades “has been about a percent and a half per year. So if you take that 1.5% and you add it up, that's actually a pretty big increase.”

Interestingly, Acemoglu also made some rough calculations of the potential impact of generative AI on the economy in the near future. And despite using a different method to calculate potential productivity gains from AI, his estimates are actually quite similar to this recent work by Deming and colleagues.

However, Acemoglu and Deming put different spins on the results. Asimoglu seems to be reacting more to all the media and industry hype around generative AI — and, he thinks, we're by no means witnessing an economic revolution. Deming looked at potential productivity gains and emphasized that it looks like generative AI could have a meaningful impact on the US economy.

“Will it lead to a 7% productivity increase? No, probably not,” Deming said. “It's not the way it's currently being used, but it can add up, and every little bit counts. Millions of dollars like that in additional GDP growth and increased living standards. It really counts.”

Deming and his colleagues plan to continue their survey in the future. It's worth noting that they conducted their last survey before Apple released an iPhone that had ChatGPT integrated into it.

Soon, he says, “as AI will be embedded in so many things, like the iPhone, it will be hard to even ask people, 'Do you use AI?'