NPR's 'Planet Money' creates an episode using artificial intelligence
A MARTÍNEZ, HOST:
The new wave of artificial intelligence technologies may be as disruptive to the economy as the steam-powered engine. Or will it be more like electricity or maybe the rise of computers? Our Planet Money podcast has been thinking about AI and a different moment in history. Jeff Guo, Kenny Malone, and NPR's GPT reporter brings us the story.
KENNY MALONE, BYLINE: One of the most common jobs for women in the United States used to be the telephone operator, the person who physically connected one phone call to another.
JEFF GUO, BYLINE: But by the early 1900s, a machine had started to replace those jobs. Eventually, hundreds of thousands of people had to find new work.
MALONE: And, you know, instead of us just telling you the story of that massive technological disruption 100 years ago, we wanted to see if today's massively disruptive technology, artificial intelligence, could tell you that story instead and possibly take our jobs in the process.
GUO: Now, if you haven't played around with these new AI tools like ChatGPT - which, full disclosure, we got free access to for our project - the way it works is you type in some instructions, and these programs sort of magically start creating for you.
MALONE: They can write term papers or legal documents or computer code. Or, in our case, we told one of these to write us a radio script about call operators losing their jobs, and it starts typing back.
AI-GENERATED VOICE: Hello, and welcome to Planet Money.
MALONE: We added the voice there, but this is what it typed.
AI-GENERATED VOICE: Today, we're journeying back to the golden age of telephone switchboard operators - when an army of fast-fingered, quick-witted humans played musical chairs with phone lines, connecting voices across towns, cities and eventually continents. To understand the impact of automation on the call center industry, we spoke with historian and author Dr. Sarah Roberts.
MALONE: Dr. Sarah Roberts? Who's the AI talking about here? - we thought.
GUO: So there is a professor of jazz in Texas by that name.
MALONE: There's a dentist in Georgia - probably not that person.
GUO: But if we had to guess...
SARAH T ROBERTS: Sure. My name is Sarah T. Roberts.
MALONE: It is this Dr. Sarah Roberts, a professor of information studies in California.
ROBERTS: I'm an associate professor and the faculty director of the Center for Critical Internet Inquiry at UCLA. And I am not an expert on telephone operators.
MALONE: Yeah, not this Dr. Sarah Roberts either. It seems the AI invented a new Dr. Sarah Roberts.
GUO: And this is a fairly well-documented problem. GPT is a language prediction tool. It's basically fancy autocomplete. So it babbles. It makes stuff up.
MALONE: So we went out and found two not-made-up experts in telephone operator automation.
DAN GROSS: I'm Dan Gross, a professor at Duke University's Fuqua School of Business.
JAMES FEIGENBAUM: I'm James Feigenbaum. I'm a professor in the economics department at Boston University.
MALONE: Dan and James are the authors of a 52-page paper on what happened when automation came for those telephone operators. And actually, to fix the AI's making-things-up problem, we had fed their paper into the computer.
GUO: And now we were going to run a little secret experiment on the paper's authors. We'd ask the AI to use their paper to write some interview questions for us. And now, we were just going to read these AI-generated questions.
MALONE: Well, let me just start. What motivated you to study the automation of telephone operation in the early 20th century? How did you collect and analyze the data for this project?
GROSS: Well, when we kind of look over the span of history and we think about what are some examples of episodes...
GUO: The interview went surprisingly normally, except for when we got to the final AI-generated question.
MALONE: It was this question about the future of automation. Dan Gross starts talking about today's automation from artificial intelligence, but he specifically reassures us, Kenny and Jeff, that AI isn't coming for our jobs.
GROSS: That AI won't be able to ask a question as incisive as the one you just came up with, so perhaps you're safe.
MALONE: Wait. Hold on. I got to - we were going to - I have to jump in. Are you aware of what's happening 'cause AI 100% generated all of the questions that we asked you?
GROSS: I don't know if we're joking anymore or not.
MALONE: We are not joking. Every question we read was generated by GPT.
GROSS: You've got to be kidding me. Why are we talking to you then?
MALONE: I mean, this is a reasonable question. This is a reasonable question.
GUO: So the AI had read an academic paper, interviewed the two authors. But could it turn all of this into a short radio story?
MALONE: Now, this part was not exactly plug-and-play. It took a lot of trial and error to get the AI to write like an NPR reporter. But what you're about to hear is completely written by AI. It even told us which quotes to use from the two economists.
GUO: It starts with the following introduction that the AI also wrote.
MALONE: I will read it now.
(Reading) In the early 1900s, telephone operators played a key role in daily life, but their jobs disappeared due to automation. A recent study investigated how this affected their lives and what it could mean for today's labor market. NPR's GPT reporter has more.
AI-GENERATED VOICE: The study by professors Dan Gross and James Feigenbaum looks at the impact of automation on young women working in telephone operation. Gross explains that automation reduced the number and share of operators by 50- to 80%, but surprisingly did not affect their overall employment rates.
GROSS: So it's not that future generations weren't able to find work when this job went away. But rather, we see them finding work in other occupations and other parts of these local economies.
AI-GENERATED VOICE: Feigenbaum adds that older, incumbent telephone operators were most affected. They had greater difficulty adjusting and often got displaced out of the labor market after automation.
FEIGENBAUM: Time to adjust really, really matters. The more time you have to adjust to the shock, the better off you are.
AI-GENERATED VOICE: The study suggests that while automation can take jobs away, the ability to adapt to the changing labor market is crucial. For telephone operators, it has led them into occupations like secretarial work and restaurant jobs.
GPT reporter, NPR News.
GUO: So there you go - a radio story about telephone operators losing their jobs to machines, completely written and reported by a machine - with, of course, a lot of guidance from us humans, at least for now.
MALONE: And if you'd like to hear more, the Planet Money team has been fretting about the future of labor in a three-part series that is out now.
AI-GENERATED VOICE: Available wherever you get your podcasts.
MALONE: Kenny Malone.
GUO: Jeff Guo.
AI-GENERATED VOICE: NPR News.
(SOUNDBITE OF KETTEL'S "QUICKPIG") Transcript provided by NPR, Copyright NPR.