AMNA NAWAZ: The last year has seen headlines of hope and outright fear around the rapid development of artificial intelligence, or A.I.

It leads to plenty of questions, including this: Does innovation lead to broad gains for everyone?

Economics correspondent Paul Solman focused on the connections between technology and prosperity with one of the co-authors of a recent book.

PAUL SOLMAN: Spot, the wonder dog, using A.I.

to navigate tricky terrain.

Already in service today at construction and manufacturing sites.

And tomorrow?

SIMON JOHNSON, Co-Author, "Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity": It could be something that helps workers be safer, helps them be more productive, and thought will get a high wage, or the same Spot could take people's jobs.

So, I think it's in the balance.

PAUL SOLMAN: In "Power and Progress," economists Simon Johnson and Daron Acemoglu surveyed the history of technological progress and came to a sobering, if familiar, conclusion.

SIMON JOHNSON: Technology changes all the time, but it doesn't necessarily turn into shared prosperity.

There's a few additional important steps that have actually been missing for a lot of human history.

PAUL SOLMAN: As in the Middle Ages.

SIMON JOHNSON: Take the medieval plow, for example, all the other improvements in agriculture in Europe more than 1,000 years ago.

What we know is that productivity increased, but there was very little change in terms of the living conditions of ordinary folk.

PAUL SOLMAN: In the Industrial Revolution, textile tech.

A power loom revolutionizing at a cost.

When the Luddites are called machine breakers, these are the machines they're breaking.

SIMON JOHNSON: Yes, exactly.

PAUL SOLMAN: Loom like this at the Boston Manufacturing Company in Waltham, Massachusetts were already displacing workers in England 200 years ago.

SIMON JOHNSON: When weaving became automated, all those people who had previously been hand-loom weavers, they got thrown out of work or they couldn't make much money.

And the Luddites were really angry, because it was the weavers who were making good money.

Those opportunities eroded and nothing else sprang to take its place.

So, like, this is not good for us, and they were right about that.

PAUL SOLMAN: Workers ditched, business owners enriched.

SIMON JOHNSON: We have a steam engine back here.

PAUL SOLMAN: On the other hand, sometimes technological progress did lead to shared prosperity.

SIMON JOHNSON: This would turn these wheels, drive the belts and transfer the power throughout the whole system.

And then if you want to use a particular machine... PAUL SOLMAN: And so, in fits and starts, the new factories of the 19th century prompted the long process of harnessing technology.

SIMON JOHNSON: They offered jobs to people who also didn't have a lot of education.

They were more productive.

Demand for labor goes up, wages go up, and trade unions show up and say, hey, how about we pay an extra wage or people have the weekend off?

Put that together, yes, you have got shared prosperity in the 20th century.

PAUL SOLMAN: Labor organizations, Johnson argues, were key to ensuring everybody shared in productivity gains.

And so it remained for almost a century.

But since about 1980, automation has outpaced the creation of shared prosperity jobs, says Johnson.

SIMON JOHNSON: People with a lot of education have done well.

People with not that much education still have jobs, but the wages at the low end have not gone up, and we're missing now that middle.

And we have been missing it for 40 years.

So it's not an overnight phenomenon.

PAUL SOLMAN: Are we in a situation in which the trend you're so worried about will reverse?

SIMON JOHNSON: We have a lot of income inequality, wage inequality, and unions in the private sector are weak.

They have had a little resurgence since COVID, but not really very much.

So, into this environment we now have A.I.

arriving.

A.I.

could be a tool to rebuild the middle class or it could be a way in which the middle class, the remains of it get hollowed out further.

PAUL SOLMAN: These are not real people.

Artificial intelligence, today's technology frontier.

A 2022 research study found that A.I.-generated faces are difficult to distinguish from human faces and are even considered more trustworthy.

At the Exploring A.I.

exhibit at Boston's Museum of Science, a car programmed by Toyota Research Institute scientists to assist drivers.

SIMON JOHNSON: The vision is not to replace human drivers.

The vision is to make humans safer, better drivers.

We do know there's 42,000 fatalities on the roads every year in the United States and almost all of those are due to some form of human error.

PAUL SOLMAN: So the hope is that A.I.

can reduce human error.

The concern, that A.I.

will instead replace jobs, like taxi and truck drivers.

Or take this retail robot, for example.

SIMON JOHNSON: So, this is an A.I.-enabled robot from Badger Technologies, goes through the store aisles looking to verify the prices and ensure that everything is properly stocked.

PAUL SOLMAN: So this is good for jobs or bad for jobs?

SIMON JOHNSON: Well, I think this one's bad for jobs, honestly, because you would previously have a whole set of people working the night shift restocking the shelves.

PAUL SOLMAN: Suddenly laid off, they'd be competing for other lower-skill jobs, lowering wages.

Same thing for drive-through robots.

SIMON JOHNSON: There are fast-food chains that have announced that they plan to replace all the human interaction.

So that entire order placing and order communication in the kitchen becomes automated by a form of A.I.

That's a lot of jobs.

PAUL SOLMAN: And plenty of much-better-paying jobs are also under threat from A.I., even those of us who may have figured we were safe.

Johnson first used the Chrome browser.

SIMON JOHNSON: We can search for you, Paul Solman, and we can get an A.I.-powered overview for the search.

Are you sure you want to go there?

PAUL SOLMAN: Yes, I'm fine.

SIMON JOHNSON: All right.

You're American, you have brown hair.

OK, we will pass on that.

(LAUGHTER) PAUL SOLMAN: Once upon a time.

SIMON JOHNSON: Build, fit.

I mean, we could just stop right there.

And you have blue eyes.

Well... PAUL SOLMAN: I don't have blue eyes.

SIMON JOHNSON: Well, you probably lied to the DMV once.

(LAUGHTER) PAUL SOLMAN: OK, not exactly reliable, but here's ChatGPT.

SIMON JOHNSON: Write a script for Paul Solman of the "PBS NewsHour."

PAUL SOLMAN: A script for this very story.

SIMON JOHNSON: Here's a suggested script.

"Good evening.

I am Paul Solman reporting on a compelling new analysis that's stirring debate in economic circles."

(LAUGHTER) SIMON JOHNSON: It's good so far.

"Cut to a visual of the book cover or relevant imagery.

The work 'Power and Progress' by economists Daron Acemoglu and Simon Johnson challenges conventional wisdom on the relationship between political power and economic development.'"

PAUL SOLMAN: And so it does.

SIMON JOHNSON: It's not bad for a first pass.

(LAUGHTER) SIMON JOHNSON: I would have put in a bit more about technology, personally, and some steam engines and a little bit of an industrial museum ambiance.

PAUL SOLMAN: Well, that's where we have added.

SIMON JOHNSON: Right, that's the human touch.

PAUL SOLMAN: And that's what Johnson is pushing, our power to harness the progress of technology in the service of shared prosperity.

What policies do you put in place to get Spot to augment people's labor, as opposed to replace it?

SIMON JOHNSON: Just like this kind of device was really Spot by DARPA, the research -- advanced research arm of the Department of Defense, when they were pushing for autonomous vehicles by having some grand challenges, you could have grand challenges where you challenge people to find ways to use these kinds of robots or A.I.

in general to develop technologies that are useful for teachers, nurses, plumbers, electricians.

PAUL SOLMAN: As Toyota is doing, for example, but economy-wide.

SIMON JOHNSON: We have done it many times, World War II, after Sputnik, the Internet, Human Genome Project, COVID vaccines.

The list goes on in terms of technologies deliberately developed with government money and with social impetus, let's say.

So this is another task like that one.

PAUL SOLMAN: And with that thought, this story comes to a close, for which I might as well use ChatGPT's conclusion.

SIMON JOHNSON: "This has been Paul Solman reporting for the 'PBS NewsHour.'

Back to you in the studio."

(LAUGHTER) AMNA NAWAZ: Please, as if a robot could ever replace the great Paul Solman.

You can watch more of our stories about the future of artificial intelligence online at PBS.org/NewsHour.