One of the main obstacles to economic growth in Chile is the decline in productivity over the past few decades, which has hindered the capacity for GDP expansion. However, this situation is not unique to our country; it is a global problem affecting developed nations as well, where artificial intelligence is mentioned as a possible solution. In this context, Chad Syverson, an academic from the University of Chicago's Booth School of Business and one of the world's leading experts on productivity, visited Chile this week to participate in the discussion "Productivity as an Engine of Economic Growth: Challenges for Chile," organized by the Institute of Economic Policies at Andrés Bello University.

In an interview with Pulso, the academic explains what has happened to productivity, the potential impact of AI, and the role of governments. Productivity has slowed down worldwide in recent decades, including in the United States and Chile. What is the reason for this?

- I wish we knew. It remains a mystery. I don't think anyone has a definitive idea or irrefutable proof to point to.

The late 1990s and early 2000s were a period when companies really discovered how to leverage the first wave of information and communication technologies. They reaped the fruits that were easiest to reach, and once that process was over, we returned to a slower productivity trajectory. But that doesn't explain why those fruits only lasted for a time.

We don't know for sure why it slowed down exactly; we do know more or less what accelerated it. The lack of productivit…

Is there a shift of production and productivity from the West to countries like China? - There has been a shift in activity, and that is undoubtedly true. A large part of manufacturing moved to China; the Chinese manufacturing industry was very productive and became more so each year.

And that shift brought many efficiency gains. But even in China, a slowdown in productivity growth has been observed, roughly at the same time. Is there a problem with the public policies of countries that have yet to find solutions?

- Well, look, if we had a dial that said "productivity growth" and we could turn it up, everyone would want to turn it up. But the problem is that productivity growth is so complex, involving so many decisions, institutions, laws, companies, and so many things interacting with each other. It is not enough to declare: "I want productivity to rise," and have that happen.

No one can decide alone to boost productivity with their individual actions. Trying to solve it really involves taking many small steps, rather than one big correct step, which makes it difficult. It has always been said that issues like education and workforce training were key to productivity, but today we live in an era with higher levels of education and still see this slowdown.

Was that diagnosis wrong? - We need to distinguish between the level of productivity and the growth of productivity. More education will increase your level of productivity, so if you go from having less education to now having a higher average level of education, your productivity should rise.

But what we are talking about is that the rate of growth of productivity is what is slowing down. And there is a limit to how much education people can receive, so to speak. You could force everyone to keep studying until they are 55, but then they wouldn't have much time to work and couldn't take advantage of that increase in productivity.

At some point, that will reach a limit, and productivity growth will have to come from elsewhere. Now, it is likely that the United States has not reached that limit. We could probably still achieve more productivity growth with more education.

So what role should governments play in improving productivity? - Many things. The basics are the rule of law, market economy, private property, those things.

But all middle and high-income countries have already somewhat resolved that issue. Now, after that, we need to talk about several other things. One of them is the incentives to accumulate productive capital, whether in the realm of human capital—meaning education—or in physical capital, which means that people with good ideas should be able to invest in them to back them with capital and make them a reality.

Should we facilitate market interactions? - Indeed, another set of policies basically refers to how much we want to improve the functioning of markets. Now, what does that mean?

It seeks a lot of flexibility, dynamism, and competition in the markets on both sides of the production process. How so? - In product markets, you know, when people go to buy goods and services they want to consume, but also in input markets, when companies want to hire workers or borrow to buy capital and build things, you want the markets to work well.

What you want to happen in the end is for the most efficient producers to grow, and the less efficient producers to shrink. People with good ideas should be able to enter the market and bring those ideas to market. And companies with bad ideas, or at least with outdated ideas, should exit the market.

In factor markets, you want workers to be able to move easily from one company to another. And you want capital to be able to move from one to another as well. Is regulation also relevant?

- In the product market, we think about antitrust legislation and other competition laws. But also labor legislation, financial sector legislation. All those things affect how well this kind of reallocation and dynamism process works.

Why is it so important to promote dynamism and flexibility in markets? - Because no one can predict what people will want tomorrow or what things will become more expensive or cheaper tomorrow. And that's why, whatever happens, you want the economy to be able to respond to it flexibly.

What we have seen is that many times governments and free competition regulators are more concerned about the size of the players than about the flexibility of the markets. Do you see it that way? - That's a good observation.

Under equal conditions, a larger number of companies in a market tends to make it more competitive. But the number of companies in the market is not the definitive measure of how competitive a market is. You can have a very competitive duopoly, and you can have a very uncompetitive market with 15 companies.

There are many other dimensions of competition beyond just the number of participants in the market. So when you think about wanting competitive markets, you need to have a very comprehensive view of what "competitive" means. For me, the key indicator is how easy it is to change suppliers.

If I don't like the version that this company offers, how easy is it for me to buy the version from another? If it is easy, it is a fairly competitive market. Is the size of firms not a problem then?

- When markets function as we said, what can happen is that if everyone goes and buys from the best company, that company becomes really large and the market starts to concentrate. But concentration is not a sign that the market is not competitive. In fact, in this case, it is a sign that the market is competitive.

The good company was rewarded with all that market share. So that is also a nuance that people should think about when considering competition. What is the role of taxes in improving productivity?

- The general rule is that we do not want taxes that are very distorting. In fact, California is trying to impose a tax on billionaires. If you earn above a certain amount, we will tax you a percentage of your wealth.

Well, guess what billionaires will do: they will leave. That tends to be an ineffective tax. In fact, in that state, they tried to avoid it by making a retroactive tax, but billionaires simply moved before it was implemented.

What impact does artificial intelligence have on productivity today? - If you look closely, you could argue that there has been a slight acceleration in aggregate productivity, and it seems to come from AI. But I think there is a lot of inherent uncertainty in that statement.

Everything points to AI starting to have an aggregate effect. But, to be honest, that aggregate data is very noisy, it varies a lot for weak reasons. So it will still take some time before we can confidently say that AI is appearing in productivity statistics.

The main fears today about AI point to job destruction and its effects on the economy. The now-famous Citrini Research study in February, which showed a crisis in 2028 as a result of this, caused markets to fall. Do you agree with this view?

- General-purpose technologies, including AI, have spread throughout the economy before. The first industrial revolution, the steam engine, electrification, the early versions of information and communication technology. All of them are general-purpose technologies.

Even agricultural production methods. In each of those cases, in the long run, they had enormous effects on certain types of employment. In the U.

S. , 170 years ago, most workers were farmers; now they are, at most, 1%. Well, is there a lot of unemployment in the U.

S.? No, not more than there was 170 years ago. The same happened with the steam engine, the electric motor, the internal combustion engine, and information technology.

I still don't see signs that AI will be different. But aren't there already some effects on jobs? - According to early indications about AI, it does not seem that it will replace people or jobs as a whole.

It seems to be a substitute for specific tasks that people perform as part of their work. The way people have used it so far is basically for those tasks, which frees them to work on other tasks that are part of their job. AI, at least so far, has been more of a complement to people than a substitute for them.

But I am not saying that "no one" will be replaced by AI. Every technology, sooner or later, will affect some type of employment. But I think we are far from thinking that most jobs will be replaced by AI.

What will happen, instead, is that many jobs will change thanks to AI. The expectations of those promoting AI are precisely that it will have implications in many areas... - Something will happen.

But this only serves to demonstrate that advocates of a technology always tend to exaggerate the concrete things that the technology does really well, and forget that, in an economy, those things must integrate with a million other things for which the technology is not the perfect solution, and it is that whole process that slows down diffusion and, in a way, mitigates any direct effect of the technology on unemployment. Now, will it take AI 30 years to diffuse as electric motors did? Maybe not, maybe it will be faster.

But it is more likely to be years. Are you optimistic about AI? - I think we should be cautiously optimistic.

As with any technology, history shows that we are usually able to manage these general-purpose technologies. It is still not obvious that it will be so fast that it will cause changes that we cannot manage in other ways. I am a bit more concerned that it could be used for malicious purposes, but that is a problem that all new technologies have.

And I suppose I am more worried about that aspect than about mass unemployment. What position should governments take regarding this technology? Should there be stricter regulation, as proposed in Europe?

- Technology is changing very quickly. Whatever is drafted to try to regulate AI will be outdated by the time it becomes law anyway. So I think it is too early.

And yes, maybe some specific problems that can be imagined today can be avoided, but benefits that cannot yet be imagined will also be destroyed. That would be a major concern. So, if you are worried, as I am, that there are actors who use AI to harm people, we already have laws against harming people.

Therefore, for me, it just doesn't make much sense to regulate AI at this moment. That doesn't mean that people shouldn't think about how this technology could be used, what kind of impact it will have on our society. Is there still a need to analyze its global impact?

- We should all be reflecting on these issues and not just take for granted that everything will turn out fine. But even if you think you could identify specific aspects that you would want to regulate, my impression is that the technology will have already surpassed that point by the time you put it in writing. There have already been some ethical and legal debates, like the recent conflict between the U.

S. Department of Defense and an AI company... And those are important conversations to keep having.

And we should talk about how to manage those issues. I just don't think that writing rules that become law is the optimal way to address those problems at this moment.