Was this text written by ChatGPT artificial intelligence? : The Indicator from Planet Money A chatbot that can do your homework for you? Sounds like every student's dream. But one college senior is building an app to call out the fakes.

For more HUMAN WRITTEN content like this, subscribe to Planet Money's weekly newsletter.

Did AI write this headline?

  • Download
  • <iframe src="https://www.npr.org/player/embed/1149834091/1149887579" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript

SYLVIE DOUGLIS, BYLINE: NPR.

(SOUNDBITE OF DROP ELECTRIC SONG, "WAKING UP TO THE FIRE")

GREG ROSALSKY, HOST:

What if in high school, instead of spending hours working on an essay about the American Revolution or - I don't know - like, "Catcher In The Rye," you could just have a machine do it?

EMMA PEASLEE, HOST:

It's basically every kid's dream. And now there's a new artificial intelligence chatbot that's making that dream a reality.

ROSALSKY: You've probably heard about this technology. It's called ChatGPT, and it could do more than just do your homework. It could write code for a website, formulate a movie pitch, help you with your mental health, even haggle with Comcast to lower your bill.

PEASLEE: All you have to do is give it a prompt like write me a college essay about the Civil War. And within seconds, boom, you have a somewhat decent essay about the Civil War.

ROSALSKY: And while kids around the world may be celebrating, like, the end of homework, this technology raises some serious questions for our education system. Like, what is the point of learning to write essays at school when AI can now do that for us? I'm Greg Rosalsky.

PEASLEE: And I'm Emma Peaslee. And we're taking over the show to talk about a new technology that not too long ago might have seemed like science fiction.

ROSALSKY: Today on THE INDICATOR - how the AI revolution could reshape everything from education to how we communicate with each other and the story of one college kid who has developed a tool that maybe makes our AI future just a little less bleak.

(SOUNDBITE OF CHRIS JONES SONG, "NO ESCAPE")

ROSALSKY: While many college kids were having fun on winter break, 22-year-old Edward Tian was feverishly working on a new app, an app to combat misuse of a powerful new artificial intelligence tool called ChatGPT.

EDWARD TIAN: I think we're absolutely at, like, an inflection point. This technology is incredible, and I do believe it is, like, the future. But at the same time, it's like we're opening a Pandora's box. And we needed safeguards to basically adopt it responsibly.

PEASLEE: Most of us have probably heard of this by now. ChatGPT is an interactive chatbot that was released in late November by a company called OpenAI. The chatbot is powered by machine learning. And I have to say, when I first heard of it, I was impressed, but also kind of freaked out.

ROSALSKY: Edward had a similar reaction. And he's sort of the perfect person to take on this issue. He studies computer science and journalism at Princeton, and he's been researching how to identify text written by AI systems. But even when he first used ChatGPT, he was like, whoa, hello.

TIAN: Me and a lot of my college friends were just asking it to write poems about each other, and we were like - going like, write a rap about someone and write a poem about someone. And it was like, wow, the results are pretty good.

PEASLEE: Because ChatGPT is more user-friendly than past AI systems and it's free, it's been able to break through in a way that other AI systems haven't been able to.

TIAN: Every teacher, like, every student - everybody was talking about ChatGPT on campus. So there was a sense of, like, oh, wow, this is everywhere.

PEASLEE: And students kind of immediately realized, whoa, this thing can do a lot of homework for us. In early December, The Atlantic magazine even declared that the college essay is dead.

ROSALSKY: Pour one out for the college essay. So, yeah, the fall semester ends, and Edward travels home to Toronto for the holidays as the buzz around ChatGPT is really exploding. And at first, he does the typical winter break stuff. He hung out with his family. He watched "Glass Onion" on Netflix. But he couldn't stop thinking about this crazy new technology.

PEASLEE: A crazy new technology that can make it hard to figure out whether something has been written by a human or not. And Edward thinks this is a problem not just for, like, kids copying and pasting their homework, but also things like propaganda generated by nefarious actors.

TIAN: Humans deserve to know when something is written by a human or written by a machine.

PEASLEE: So on holiday break, Edward decides to create an app to try and address that.

TIAN: I just had so much free time over winter break, and I was getting really bored. So I was like, wow, why don't I just code this out so the world can actually use it?

ROSALSKY: I mean, we've all been there. Why not just create an app over winter break?

PEASLEE: Totally. So on January 2, Edward released his app. He named it GPTZero. The app basically uses ChatGPT against itself. And he was able to do this because he could tap into the data of an earlier open source version. So when GPTZero analyzes a text, it can scan to see if it recognizes any of the same patterns that ChatGPT would generate.

TIAN: So we're basically taking one of these, like, text generation models and asking it, hey, is this new piece of text - does it seem, like, pretty familiar to you? Like, would you probably generate it yourself?

PEASLEE: When Edward went to bed that night, he didn't expect much from the app. But the next morning, his phone had blown up. He had so many texts and DMs from journalists, principals, teachers, you name it, from places as far away as France and Switzerland. It became so popular that the app crashed.

ROSALSKY: Before all this, Edward's biggest plans were graduating from college and getting his wisdom teeth pulled. Now he's fielding calls from venture capital firms, education leaders and global media outlets.

TIAN: My own high school English teacher reached out to me. My, like, principal reached out. They're telling me, like, they're in, like, a big group chat of, like, high school principals and teachers that are all trying to talk through how to handle ChatGPT.

PEASLEE: But not everyone thinks ChatGPT is a problem. Some prominent techies are even celebrating it as the end of homework. And a lot of this talk is probably hype. But it really does feel like we've entered a new world where we're being forced to reevaluate our education system and even the value of teaching kids how to write.

ROSALSKY: But, you know, like, some of you out there may be too young to remember a time before everyone had cellphones. But I remember those times. And because of that, I remember having to memorize phone numbers. Now I don't have to do that, and I don't do that.

PEASLEE: And a techno-optimist might be like, yeah, that's for the best. It frees our minds to concentrate on other matters, but not that helpful if you, say, lose your phone and need to call someone.

ROSALSKY: Now with ChatGPT, it's possible to imagine, like, this dystopian future where all written communication is written by a machine, a world where everything we write to each other is like a Hallmark card written without our, you know, personality or ideas or emotions.

PEASLEE: But at least when you give people Hallmark cards, they know it's from Hallmark. If you use ChatGPT to write your friend a congratulations or an apology, they might not even know it was written by a machine.

ROSALSKY: Which brings us to the other purpose that Edward envisions for his app - to identify and incentivize originality in human writing.

TIAN: There is, like, a celebration of individuality in writing, and we're losing that individuality if we stop teaching writing at schools. There are aspects and beauty in human writing that computers never and should never co-opt. And it feels like that might be at risk if everybody is using ChatGPT to write.

PEASLEE: But Edward is no Luddite. He's not trying to, like, stop AI in its tracks. He actually opposes blanket bans of the technology like the recent bans by New York City and Seattle Public Schools.

TIAN: Banning it is not going to stop the wave of AI that's inevitably going to come. Like, even if you ban the ChatGPT usage on school Wi-Fis, kids might use it at home. It's like you can't stop something that's going to be inevitable. So we can't enter this future blindly.

ROSALSKY: As for his plans after college, Edward says the excitement and the clear demand for his new app have convinced him that he should concentrate on making it better and more accurate. Like, he's thinking about pursuing this full time.

TIAN: Top priority is building this out into something that teachers can use, like, day to day in their workflow. So if you're a teacher or an educator, I'd love to talk to you. Our team, which is, right now, just me and my best friend from college, who just joined yesterday, would love to talk to you.

(SOUNDBITE OF MUSIC)

PEASLEE: So if I ever get a thoughtful card from you, Greg, I'm probably going to run it through ChatGPTZero (ph).

ROSALSKY: (Imitating robot) You are my favorite colleague, Emma. This is not written by a machine.

(SOUNDBITE OF CHRIS JONES SONG, "NO ESCAPE")

PEASLEE: This episode was produced by Brittany Cronin and engineered by Debbie Daughtry. Sierra Juarez checks the facts. Viet Le is our senior producer. Kate Concannon edits the show. And THE INDICATOR is a production of NPR.

(SOUNDBITE OF CHRIS JONES SONG, "NO ESCAPE")

Copyright © 2023 NPR. All rights reserved. Visit our website terms of use and permissions pages at www.npr.org for further information.

NPR transcripts are created on a rush deadline by an NPR contractor. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.