One area where artificial intelligence has been swiftly adopted is software coding. Google even boasted last year that more than a quarter of its code was generated by AI.

But the technology is also generating challenges to the traditional technical job interview, where candidates are given programming problems as a way to assess their skills. And lately it’s become apparent that a lot of applicants are using AI to give themselves a boost, according to recent reporting from Business Insider's Amanda Hoover.

Marketplace’s Meghan McCarty Carino spoke with Hoover about the controversy over applicants using AI while interviewing for jobs that often use AI.

The following is an edited transcript of their conversation.

Amanda Hoover: They're on a video call with somebody who's supposed to be coding, and the person might be really looking off to the side a bunch. If they also see like blocks of code appearing, that is a good hint that it's being copied and pasted because if you think about like if you're watching somebody code live, you would be watching it be typed out, watching them maybe work back and forth to solve the problem. And then additionally, there were verbal answers, you know, given like questions asked and they're just reading off something that's not related to the question.

Meghan McCarty Carino: So what are hiring managers that you spoke with doing to respond to this?

Hoover: Some just watching for these signs and maybe blacklisting people who appear to be cheating. Others are doing things more so where they're allowing people to use [generative] AI, you know, in the interview process, because part of this is that it's almost expected in many companies now that engineers are going to use AI on the job. If that's how you're going to do your job and how you're going to show up each day and be efficient and solve problems, you know, because you still need to understand everything that's happening, even when you're using AI, why can't you use it in the test? Like, that's a question a lot of people are having. Some people or some companies are really going to be reevaluating their tests because I spoke to some experts who are saying, you know, if you're going to invest all your time into trying to catch cheaters and enforce these rules, like, that might not be the best use of your time. Instead, it's finding ways to interview better and really have deep conversations with applicants to see what they understand.

McCarty Carino: Right. It seems like a fundamental question that you're bringing up is should it even be considered cheating to use AI on these tests?

Hoover: Yeah, I think there's a lot of frustration in the job market right now. Ever since tech companies started laying massive amounts of people off a couple of years ago, you know, there's been so many applicants for every job, and that's coincided with the rise of generative AI. So imagine, you know, you're applying for hundreds of jobs to try to get interviews. It's really overwhelming for applicants, so some of them are turning to AI to, like, write their cover letters, edit their resumes, mass apply to jobs. These coding tests are unpaid labor, and some job applicants really feel taken advantage of, so they think if the recruiters are using AI, if AI is kind of part of the job, like, what's the big deal to use some of it here to become more efficient in my job application process? There is, obviously, there's an ethical problem of telling, or, like, agreeing to not use it and then going and doing that. Like, you're lying to your future employer, that's not a great way to start out. But I did speak to people who were saying, like, perhaps a tipping point is coming. You know, if 90% of people are going to use it in some way, that's so unfair to the few people that are trying to, like, really follow the rules superclosely, and, you know, not have that assistance. So I think we're gonna see a change in this conversation about what counts as cheating in these situations.

McCarty Carino: One interesting wrinkle you bring up is the issue of kind of younger workers who are coming up in this environment where maybe they're using, you know, an AI assistant for coding from the get-go. Does that result in a certain amount of "de-skilling" that employers kind of need to be able to test for? Tell me more about kind of the challenges that this poses.

Hoover: Yeah, it's something that employers are a little bit concerned about. You also have companies that are more and more using AI to do the work of junior engineers, and maybe hiring less or using more AI to write more code. And it's raising big questions about, what is the career ladder going to be to come into this industry if you have more senior engineers working with AI instead of working with junior engineers? So it's also raising concerns, you know, if they're going through college and professors aren't finding new ways to evaluate students, to really focus on if they're using AI or how they're using AI. You know, they might ace all of their classes and then get into a working environment or a test like this, and it might be less clear than in the past how much they can really do on their own, or how much they understand all of it. That's some concerns I've heard from people that have been in the industry a bit longer.

McCarty Carino: Right, because often we talk about people, sort of, you know, being the boss of the AI. So the AI kind of does the grunt work, but if the, you know, the person using it as an assistant has more mastery, has kind of the broader context, they can kind of see where the problems arise or little things that need to be fixed. If you've always kind of come up with an AI assistant, maybe you can't be its boss?

Hoover: Yeah, yeah. I mean, I think that people from the outside of this industry really look at coding as, like, so scientific, but there's a lot of creativity involved in it as well. When I've talked to people who have these concerns, they do really say that creative developers are going to still, you know, have jobs and make great products and write great code. The shift might go away a bit from technical skill and more towards creative problem-solving becoming of higher value than it already is.

McCarty Carino: What do you think this kind of situation says about the future of hiring, both for employers and for job seekers?

Hoover: Yeah, it's really becoming complicated …outside of the engineering world, outside of the tech world even. One of the biggest pain points is that people can just mass apply and recruiters are just really feeling overwhelmed and they might be using AI to try to sort through these applicants. That raises all these issues where we, you know, don't know all of AI's biases. We do know that in the past, it's had problems being biased racially or by gender, and you definitely don't want to see people not being considered for jobs that they might be great fits for if the AI makes those mistakes. So it's really creating a lot of confusion and frustration on both ends, and a little bit of a breakdown happening. It's a problem that's going to take, I think, multiple solutions to get us back to, you know, getting the best people connected to the best jobs and without all of this noise of the extra applicants going on.

More on this

Perhaps one of the more infamous coders who admitted to using AI is Chungin "Roy" Lee, a former Columbia University student who created an AI tool that supposedly helps job applicants cheat on their coding interviews.

Lee claimed — quite publicly — it helped him get internship offers from Amazon, TikTok, Capital One and Meta.

Lee was suspended from Columbia for sharing a recording of a disciplinary hearing over that same tool but he's leveraged the publicity to build the brand of his startup, called Interview Coder. He told NBC that he's now making almost $200,000 in monthly subscriptions.