Ask AI
Back Back
MedBuzz: FDA Shake-Ups, Student Loan Policy, and AI Hallucinations
Activity Information Ask a Question

Activity

Progress
1
Course Completed
Activity Information

Released: May 27, 2026

John Marshall
John Marshall, MD
Catch Up On All Episodes of Oncology Unscripted

Watch Now

MedBuzz: FDA Shake-Ups, Medical Debt, and AI Hallucinations

John Marshall, MD: John Marshall, Oncology Unscripted. I really wanna thank the folks at Decera for helping to support the work that we are doing, and all of our team behind the scenes who make this possible, so special shout-out to them. But there is a lot of news out there. Let’s start at the highest level.

The front page of The Washington Post this morning says that the FDA is gonna get a new leader. Marty Makary, a surgeon from Johns Hopkins, a GI surgeon who’s been filling in that spot at the head of the FDA, resigned officially this morning. And it’s just a big shake-up that’s going on in terms of staffing, in terms of leadership, in terms of priority there. And there’s really been a tension. Marty was one of those guys who was actually doing a pretty good job of balancing the pressures from the current administration and trying to do the job of the regulatory agency, of approving things that are good for people and not approving things that aren’t good for people.

But they keep relaxing the rules and changing, and basically, I think the two sides couldn’t meet. So, he’s like, “Enough. I’m outta here,” even though our president gave him a good review walking out of the door.

But, you know, there was a recent article—an approval, actually—of flavored vapes, right? We’re trying to get fewer and fewer people to smoke and not wanting young people to start these habits, but flavored vapes probably are gonna drive young people to this even more. And Marty originally didn’t want this, but ultimately caved, and they did grant approval for that through the regulatory agency. So, we’ll see what happens without the backstop of somebody like Marty in that position. That was one thing that happened this week in the news.

Another: there’s been a lot of talk about the financing of education, and there have been some recent data that the current administration wants to pull back on supporting education. And this is gonna lead to more and more student debt. And this is nowhere more obvious than in the training of physicians and other health care providers because it’s expensive, and there’s a lot of delayed gratification.

Then again, this morning in The Washington Post, there’s a paper—an article—in the finance section of The Washington Post about, you know, not only are we really good at delayed gratification when we go into health care because we have to keep putting off really making any decent money, but if you pile up a bunch of debt, you also have to put off saving money for your future because you’re paying off this big debt burden. So, it’s almost like having a second mortgage going forward.

When the government says, basically, “We’re gonna stop funding or supporting trainees in this way,” what they’re not really thinking about is the downstream impact. And we already know we have a shortage of medical and health care providers out there in this country, so increasing brain drain is gonna happen there, which is gonna have its impact 10 to 20 years from now, something that the current administration can’t see forward enough. So, financing medical school and access to support is a big deal in our world.

I promised you that there’s going to be a lot of AI talk. Even in our last episode of Oncology Unscripted, I came out and talked about how, at least in health care and science, I was increasingly comfortable with using AI to do searches. But then Maxim Topaz came out with a paper that was just recently published in The Lancet that’s been incredibly referenced all over, and he has been tracking—he and his team have been tracking—how many hallucinated references are actually then published out there in the literature.

And he showed this alarming trend where, year after year, it went from, like, 1 in 2,800 papers having one fabricated reference to now 1 in 277. So, it’s just skyrocketing, the number of papers that are published where at least one or more of the references cited is actually not a real reference.

And so, we’re going to have to be increasingly careful. Unlike what I just said a couple of weeks ago, where maybe we should be more comfortable in science, his paper suggests that maybe that was wrong, and I need to backpedal, and we need to be very careful.

Now, when you were in medical school and you were taught how to take a history and how to interpret that history, it always came to the awkwardness of: do you smoke? Do you drink? And if so, how much? And so, I don’t know if your training was like my training, but I was told and taught that whatever answer you give—how many drinks a week or a day that you have—we’re supposed to sort of secretly, in our own heads, double that, right?

That’s what you were told. Now, I tell my primary care doctor exactly the right number. Now, whether she—she’s a friend—doubles my number or not, I don’t know. But I’m actually giving the correct number that I think is in the medical record. I don’t check her work to see if she writes the number I give her or if she writes a double number.

I was also thinking about a recent story about this guy named Kash Patel. You probably know about Kash Patel. He’s leading the FBI right now, and there’s been a lot of talk within these congressional meetings lately about how much he’s been drinking.

He was filmed in a celebration in a locker room after one of the recent teams won a championship. When he was confronted with this “you drink too much” kind of thing in public, of course, he’s suing everybody for that statement. What’s ringing true to me is: do I double what he’s saying? But he’s protesting so much, it made me flash back to Hamlet.

It’s actually Act 3, Scene 2, where the line is, “The lady doth protest too much, methinks,” right? We all know the line. And when somebody’s fussing that much about something, it makes you wonder if it’s indeed maybe possibly true. So, do we double what he says? Do we cut and... I don’t know what to do with all of this.

But there’s a lot out there this week in our news that’s gonna affect us in health care today. Most of it is unsettling; most of it is troubling. But you can ask me whether you think this is true: am I protesting too much, or am I on target?

Oncology Unscripted, John Marshall. Thanks for joining me.

This transcript has been generated by AI.