Will AI Chaos Precede AI Success?

A Conversation with Greg Leo

The potential economic impact of AI — both good and bad — is as frequent a matter of analysis as is the revolutionary potential of the technology itself. Greg Leo teaches economics in the LMU Bellarmine College of Liberal Arts. He specializes in microeconomic theory and experimental economics. We spoke to him about AI’s impact on the economy, from workers to a possible bursting of an economic bubble. Leo was interviewed by Joseph Wakelee-Lynch

Job loss is considered one of the most significant negative economic effects of AI. How threatening is AI to white-collar workers?

A lot of jobs are involved in the production of content. I write textbooks, for example, and articles, and exams. I can have AI write an exam. But writing an exam is not as simple as just putting questions on a piece of paper. I have to make sure that the exam is tailored to the course difficulty, is covering the topics that I’ve covered in the classroom, has the right level of challenge but isn’t too challenging. You cannot automate the checking of whether that’s a good exam. There are a lot of things AI can do, but not in a way that’s acceptable.

What about the impact as related to age?

You can imagine that younger, entry-level workers doing more routine tasks are definitely more threatened by this. Which creates a problem because you have to have a pipeline of people who are being trained. If we’re not hiring entry-level workers and training them to do a job well, then there is going to be a generation of under-trained workers. The occupation that has been most affected is coding. That means that entry-level coders are not building skills like they used to. If we’re missing this pipeline that turns entry-level programmers into talented programmers, that’s going to lead to problems in the future. 

Do you believe that there is an over-investment in AI technology?

The idea that we’re overinvested in something is difficult to answer. With AI, we’re invested in something that may happen. At some point, if we put enough money into it, we will have created something that will replace all white-collar work so we don’t have work anymore. I would say it doesn’t appear that we’re moving in that direction as fast as we want to. If you look at investment patterns, almost all venture capital money has gone into AI. All the growth in the United States in 2025 is basically just in data centers. But it’s hard to find any measurable growth in productivity. It’s not clear that it has made jobs better. If AI were to go away today, was that worth it? If we can get to artificial super-intelligence, maybe the answer is yes. In the absence of that, we’re just going to keep pouring money into something that hasn’t paid off.

As an economist, are you concerned that we’re headed for an economic bust? 

The structure of the economy right now does not look very good. Even the CEO of OpenAI, Sam Altman, thinks we’re in a bubble. What’s going to happen if the bubble pops? The first companies to suffer are start-ups. There is nonsense money going to all kinds of start-ups. It’s the chaos part of development: You don’t quite know what’s going to work yet, so you put money into everything.

So, should we expect the bursting of the bubble to have extremely widespread economic effects?

It’s hard to speculate. The economy is tightly woven. When any big part of it breaks, there is collateral damage elsewhere. I’m probably more confident about when the bubble bursts and we get beyond the chaos stage. My prediction is what we’ll see is a shift from thinking about AI as a race for artificial super intelligence to thinking about actually productive, meaningful ways to utilize AI that aren’t super expensive, and actually make people better off. There may be a silver lining there. The sooner it happens, the sooner we can move away from the narrative of what will it do to work and the sooner we get to the narrative of what can it actually be utilized for.