Americans are using artificial intelligence at fairly high rates. What does this mean for the economy? Money Planet: NPR

Before making a Last scan Of Americans to find out how much they use generative AI, David Deming, an economist at Harvard University, says he is firmly in the “AI skeptics” camp. That is, he was skeptical that the explosion of generative AI would provide significant benefits to the American economy any time soon. However, he now says he is more optimistic.

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

Deming says the study was motivated by questions about whether and how much Americans use generative AI. In doing what economists tend to do, he and his colleagues Alexander Beck and Adam Blandin wanted to get some good data.

They modeled their survey on the Current Population Survey (CPS), which is sponsored by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics. Conducted monthly, the CPS is a kind of gold standard for surveys. It’s how we know things like the unemployment rate and the state of the labor market.

In sum, the survey conducted by Deming, Beck, and Blandin is of high quality. It has a large sample size. She is represented at the national level. They also included some of the same questions that CPS asks, so they can compare their survey with CPS and make sure their numbers are good. They conducted their survey twice, in June and August 2024.

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

What’s even more surprising, Deming says, is that the use of AI seems to be widespread across almost every field. He expected younger, more educated Americans to be the biggest users. They confirmed that this was the case. “But we found that 22% of workers say they use AI, and usage rates were higher than 20% in every major job category except personal services, where the rate was around 15%,” he adds.

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

What does this mean for the economy?

Recently Two-part series on IndicatorDarian Woods and I debated whether AI is overrated. Given the enormous uncertainty about the impact that AI might have on the economy – as well as our reluctance to make predictions that might make us laugh in a few years – we decided to withhold our personal feelings on the matter. We flipped a coin generated by artificial intelligence. I got “AI is overrated” and wrote episode And a Newsletter Make this a controversial issue.

In the course of my research, I discovered a study conducted by the Census Bureau this year that found just that about 5% Of American companies said they used artificial intelligence in the past two weeks. I used this as evidence that the use of generative AI was pathetic, especially given all the hype surrounding it.

How does Deming reconcile his results with this result?

For example, Deming says he and his colleagues found that Americans reported using AI more during their personal time at home than at work.

Ummm. Yes. When I repeat the saying “AI is overrated,” it makes me think that much of the use of AI is for fun rather than for productive work purposes — which suggests that the impact of AI on the economy will be limited.

The economic dream of AI would be that it would rescue the US economy from a long period of slow productivity growth. Productivity growth – meaning workers can produce more in less time – is the magic sauce for rising living standards. Modern technologies have been largely disappointing on this front.

I mean, look at the smartphone. If I told you back in 2006, the year before the iPhone launched, we would soon all have super-fast computers in our pockets, capable of searching the Internet, giving us precise directions to any location, sending emails and making video calls to our partners. – Workers and customers, basically ordering any product or service, translating languages ​​etc. – You would think we would see an explosion in productivity. But it seems that smartphones have become more of a tool for fun and distraction than an incredibly powerful work tool. We have not seen a huge boost in productivity growth since this policy was widely adopted more than a decade ago.

However, Deming emphasizes that people are using generative AI at work. He says their survey shows that “about 1 in 4 people have used it at least once in the last week at work” (versus “about 1 in 3” who have used it in their free time). When it comes to how people use AI to help with their work, participants indicated that writing, interpretation, and administrative support were the most helpful (although a large percentage of them also said they use it for other tasks, including programming and support when dealing with clients and colleagues the job).

Comparing his study to that of the Census Bureau, Deming says he believes there is “a big gap between official company policy on the use of AI and workers. I think a lot of the use falls under the radar, like you should write an email to your boss, and you just use it to write Email is faster, and you use ChatGPT to do it, but you don’t tell anyone. Deming says many of his students use generative AI this way.

We reached out to Daron Acemoglu, an economist at MIT—what these days might be called a leading AI skeptic—about Deming and colleagues’ study. Does the fact that they found that a large percentage of Americans use generative AI change his view at all?

“What concerns me about their numbers is that they don’t distinguish between core productive uses of generative AI and incidental/trivial uses,” Acemoglu said in an email. “If you’re curious about what ChatGPT would say to introduce one of your guests and use it to learn about that, you’re a user, but that’s not fundamentally built into NPR’s production process. Worse still, if you ask me if I’m using generative AI, and if I’m being honest, I would have to say yes, because when I do research, I now see generative outputs generated by AI, but this has no impact on me, as productivity improvements require fundamental integration and reorganization of production processes, along with complementary investments in organizational capabilities and worker skills – Not trivial and sneaky uses We know that ChatGPT has about 200 million unique monthly users, but the question is how many of them are using it in some way that will lead to significant productivity improvements/cost reductions, I don’t really know the answer to that question, I don’t think We will find the answer to this question in this paper.

To what extent will generative AI increase productivity?

To try to predict the extent to which AI could boost productivity growth, Deming and his colleagues performed some rough calculations. They looked at five randomized studies that analyzed the extent to which the use of generative AI increases productivity in different work environments. They chose the productivity increase number in the middle of those studies, which is about 25%. They then multiplied that, Deming says, by “our estimates of the number of work hours currently supported by generative AI” (which, according to their calculations, is between 0.5% and 3.5% of total work hours in the United States).

They ended up estimating that generative AI would lead to “an increase in labor productivity of between 0.125 and 0.875 percentage points at current usage levels.” That may not seem like much. But, Deming says, keep in mind that productivity growth over the past two decades “has been about one and a half percent per year. So if you take 1.5% and add that up, that’s actually a very large increase.”

Interestingly, Acemoglu also made some rough calculations of the potential impact of generative AI on the economy in the near future. Although a different method is used to calculate the potential productivity increase from AI, his estimates are actually very similar to this recent work by Deming and colleagues.

However, Acemoglu and Deming put forward different interpretations of the results. Acemoglu seems to be reacting more to all the media and industry hype around generative AI – and it seems that we are not witnessing an economic revolution by any means. Deming looks at the potential increase in productivity and further emphasizes that it appears that generative AI could have a beneficial impact on the US economy.

“Will that lead to 7% productivity growth? No, probably not,” Deming says. “It’s not the way it’s currently being used, but it can add up a little, and every little bit counts. That’s millions and millions of dollars of additional GDP growth and rising living standards. That really matters.”

Deming and his colleagues plan to continue conducting the survey in the future. It is worth noting that they conducted their last survey before Apple released an iPhone with ChatGPT built-in.

“Very soon, because AI will be integrated into so many things, like the iPhone, it will be difficult to even ask people, ‘Are you using AI?'” he says.