About 1 in 4 U.S. jobs requires an occupational license, according to the National Conference of State Legislatures. Licensing requirements differ by state and can apply to everyone from barbers to lawyers. The general idea, of course, is to keep unqualified workers out.

But technology, and specifically artificial intelligence, is making inroads. Rebecca Haw Allensworth, a law professor at Vanderbilt University, is also author of the new book “The Licensing Racket: How We Decide Who Is Allowed to Work, and Why It Goes Wrong.” She told Marketplace’s Stephanie Hughes that in some instances, AI is letting consumers bypass licensed workers altogether. The following is an edited transcript of their conversation.

Rebecca Haw Allensworth: This has been true for a while, but it’s really become intense now that AI is what it is. You can go on ChatGPT and say something like, write me a contract for funeral services, and ChatGPT will spit out a contract. That has forever been the practice of law, and that’s been limited, until this point, to people who have a license in law.

Stephanie Hughes: But now AI is kind of getting in on that practice of law, even though there have been lawsuits to try to stop it.

Allensworth: Yeah, it’s hard to know who to sue. So if you were not a lawyer, if you were just an unlicensed person, and you were writing contracts for people, the [bar association] might come after you and say, you’re engaged in the unlawful, unlicensed practice of law, and you have to stop. A cease-and-desist letter. With technology these days, it’s a little bit hard to figure out who was writing that contract. Should I get in trouble for asking the prompt? Should [ChatGPT developer] OpenAI as a company have some liability for this? And if that were the case, how on Earth would we figure out what actually is happening at the company level? So, yeah, it’s breaking down some of the barriers of licensure, and I think it’s really unclear how the professions are going to respond.

Hughes: What’s the worry here? Like, the worry is that the bots are not licensed by any particular professional board and maybe don’t know what they’re doing — is that the concern?

Allensworth: Well, I think there’s two concerns. The first concern is that somebody getting legal advice or a contract written by a bot is going to get bad advice. But the other concern, I think, is on behalf of the profession itself. If that contract is a straightforward, easy thing that can kind of be pulled together by the bot, then that’s a really inexpensive way for me to get a little bit of legal help. And lawyers don’t like that because that will, at the end of the day, undercut their bottom line.

Hughes: It seems that sometimes these professional licensure boards are doing their best to keep technology out. In the book, you write about a company called Edge AI that dealt with this. Can you tell me a little bit about that?

Allensworth: So this was a case that I saw in front of the alarm system installers board, which, believe it or not, was until recently a licensed profession in my state.

Hughes: So in Tennessee, if you wanted to have an alarm installed, you had to have a licensed person come by and do it?

Allensworth: That’s right. And in another case, I saw that same board go after somebody who had installed their neighbor’s Ring cam from Sam’s Club for unlicensed installation of an alarm system. But this case was a little bit more threatening to the board and to the profession because this guy had figured out how to make the AI to recognize faces. So he was selling this to, like, schools and hospitals as a way of keeping out known sex offenders or other people who shouldn’t be on the premises. And he wasn’t really actually installing an alarm system, exactly, but the board thought it was kind of close enough and said, well, hold on, you’re like a startup programmer, but did you also know that you are an alarm system installer, and as such, you don’t have a license? And he says this effectively shut down his business because it was going to take him months to try to get a license, something he couldn’t do. He just lost those months, and competition caught up with him, and he wasn’t able to start his business.

Hughes: I want to ask about telehealth. This really took off during the pandemic, and it seems like it sort of pushed the boundaries of how different states license medical professionals. Can you tell me about this?

Allensworth: The licensing boards had been resisting telemedicine in various ways for a long time. COVID kind of made that a little harder for them to do. They couldn’t make the same arguments that this just isn’t going to happen, this isn’t safe, because there was such a demand for it. And so you did see, because of this big shock to the system, an expansion of what licensing boards were willing to let doctors do remotely. At the same time, it’s still true that you have to be in the same state as your doctor. The doctor seeing you has to have a license in that state. So there’s another way in which, if we didn’t have such kind of state-by-state and onerous licensing laws, telehealth could be used on a national basis, really expanding access to care.

Hughes: Is there a world where licensing boards could modernize, like could accept technology faster and kind of just move things along?

Allensworth: Well, like so many things I saw at the licensing boards, this is one of the downsides of licensure that boards just don’t understand, the way in which it makes the profession slow to react to technology. And I think part of why boards don’t understand the real effect of their regulation is that they’re super-expert in their profession, but they’re not expert in the consequences of the regulation. They’re not expert in regulation. These boards are basically self-regulatory. They’re made up of members of the profession making these decisions. And I think if we had people making these decisions who thought about the importance of technological change and innovation, they could modernize in the way that you’re talking about. But I don’t think the way they’re currently constituted — mostly members of the profession — you’re going to get there.

Hughes: So you’re a licensed lawyer. How are you thinking about the balance between technology and licensure and how it affects you as a practicing lawyer?

Allensworth: Well, I mean, I think about this a lot in my role as a teacher of practicing lawyers because pretty soon it’s going to be malpractice to not be able to use these tools. The whole way in which we train lawyers is pretty clunky and slow to change, and I just think that we need to not be afraid of it. I think we need to understand it, and I think we should also be excited about it because I think this idea that you can go online and ask ChatGPT to write you a contract or ask it a legal question and get a decent answer, it’s not the end of the world. It can really expand what we call access to justice. Access to care is the same idea, but in medicine. There’s a lot of people who can’t afford legal services, and I think we should look at AI as an opportunity to expand the provision of those services. We just have to find a responsible way to do it.

Hughes: What do you see as the future, in terms of how licensing boards deal with technology or think about modernization?

Allensworth: Well, I think that some things force their hand. So I think COVID forced the hand of medical boards to really reckon with telemedicine. And I think probably generative AI is going to force law boards to really reckon with what counts as the practice of law. And I think that what you’ll see is resistance, but you will also see some change forced upon them. And I think that in the case of law, the future is a lot of things that were considered to be the practice of law, and therefore requiring a license and requiring a live person, are now going to be seen as not that and can be provided by AI. And I think that’s a good thing.