KJZZ is a service of Rio Salado College,
and Maricopa Community Colleges

Copyright © 2025 KJZZ/Rio Salado College/MCCCD
Play Live Radio
Next Up:
0:00
0:00
0:00 0:00
Available On Air Stations

ASU professor predicts AI may take the place of regular doctor visits

California health officials will start using artificial intelligence to translate certain websites and documents into a variety of languages. An official with California Health and Human Services Agency says human editors will oversee and edit the translations.

This is just one area in which AI is being used in health care. James Hodge believes AI will only become more widely used in that field going forward.

Hodge is a professor at Arizona State University's Sandra Day O’Connor College of Law and director of the Center for Public Health Law and Policy. He joined The Show to talk more about AI in public health.

Conversation highlights

James Hodge
Amber Victoria Singer/KJZZ
James Hodge in KJZZ's studio.

What do you think of this plan in California?

JAMES HODGE: Hey, listen, this is exactly what might be considered one of the most innocuous uses of AI as we go forward. The opportunity to provide real information to consumers and health care and other arenas through translatable information is a success story for AI. If it can do it accurately, Mark, and if it can't do it accurately, now, we've got some problems.

Well, that's the real question. How do we know? How do we try to keep tabs on that?

HODGE: Well, we could use humans to actually overview and assess: Was that translation into this foreign language accurate based on what we're predicting for this particular site? That's possible, of course, but that's time consuming. I expect AI will be checking its own devices and own output through different interventions to make sure the accuracy is as good as we can get it.

Does that give you pause — if A I is checking up on AI?

HODGE: Wouldn't it for you? ... If you're talking about medical information, the type of which consumers may rely upon, especially which they're going to a site and hearing it in a language that sounds familiar to them. If it makes mistakes, we better hope something's checking on it, but that check itself could be as flawed as the original information.

So what would you like to see in terms of how this might be rolled out? What would make you confident that this could be as innocuous and as helpful as some people think it is? What would be the way to roll it out in your mind?

HODGE: Well, I think a gold standard would be, of course, we can actually utilize AI fully to make the translations work within that specific environment. And vet those cleanly with humans involved. Particularly medically trained persons who can validate the data is actually an accurate and fair assessment of what persons in that culture, ethnicity would be listening to and understanding from the language is translated. But we're not gonna get that gold standard. I don't believe, on all of these translations. That's time consuming and, frankly, AI is designed to, you know, expedite that. I think AI is gonna be checking up on the accuracy across different you know, programs. That's what we're going to watch for very closely.

Are there other forms in which AI is being used in health care? I can't imagine that there aren't.

HODGE: It's almost a question mark at this point. Where is it not being used in health care delivery? Let me give you some common examples. So we're all very familiar with what — where this is now and where it's going. So in regards to artificial intelligence and its implications and repercussions for health care delivery, it is currently being used in the health insurance industry to determine who's gonna get vetted for what specific care. Doctors are using it to sort of assess patients records and information. Are they providing the most accurate care? It's being used in prescription industries, websites that would actually take in information and provide prescriptions based on what you provide to that specific site.

It's a question, not so much of what other applications, but what application is it not being used. And that's the beginning. It will only continue to proliferate extensively over decades to come so much. So that one day, I'm predicting, your main health care delivery may be through AI, and you'll see a doctor only when absolutely necessary.

What are you seeing in terms of how effective AI is in some of those things?

HODGE: Well, let's talk about one example, in regards to radiologic assessments, for example. Its accuracy there can be outstanding, and it is far more efficient than humans might be on those specific fronts, [like looking at X-rays] ... There are other types of data we may get through electronic transmissions about human tissue or otherwise. It can be really accurate, really capable of providing some of the better insights as to what the current status is for detecting everything from tumors to specific tears or other things within muscular structure.

Now, with all that said, even though it's got its great applications there that are huge improvements over what humans can do, the risks of what and how far AI could go in these arenas, particularly where you see it used as a gatekeeping program. The type of program that you're seeing already. Big insurers like Humana being sued over because their gatekeeping functions are what's telling Medicare patients right now, in some cases, you get access to this care or you don't. Based on AI — not a doctor making that assessment or what we call a utilization reviewer, but an actual program. That's happening right now as we speak.

KJZZ's The Show transcripts are created on deadline. This text is edited for length and clarity, and may not be in its final form. The authoritative record of KJZZ's programming is the audio record.

Mark Brodie is a co-host of The Show, KJZZ’s locally produced news magazine. Since starting at KJZZ in 2002, Brodie has been a host, reporter and producer, including several years covering the Arizona Legislature, based at the Capitol.
Related Content