According to a study in Nature, in 2023 alone, more than 10,000 research papers published in academic journals were retracted, which was a new record for a single year. The rate of retractions, in fact, has more than tripled over the past decade.

At the same time, those retractions have a silver lining: They mean the science community is catching fraud. One of the leaders in that effort is Dr. Elisabeth Bik, a microbiome expert and a science integrity consultant. She joined “Marketplace” host Kai Ryssdal to talk about what she’s seeing in the world of research and journal publishing. An edited transcript of their conversation is below.

Kai Ryssdal: How does a microbiologist become a science integrity person? Could you tell us how you got there?

Elisabeth Bik: Sort of a strange career switch, I guess. Yes. I discovered that somebody had plagiarized my papers, and that made me very angry. And so I worked on plagiarism for a while, and then, by accident, I discovered somebody had duplicated images that were representing different experiments. So then that became my new hobby, and now it's my profession.

Ryssdal: Well, give us the 30-second talk on how big a problem this is in the academic publishing industry, which is worth billions and billions of dollars. And as you've demonstrated, some of it's just fake.

Bik: It is. A couple of years ago, I did a big study of 20,000 papers, and I found that 4% of those papers contained duplicated or manipulated images, and we estimated about half of those — so 2% of those 20,000 — were done deliberately, so a Photoshopped image, for example, or an image that had been rotated or mirrored. But I think since then, the problem has only grown, because there's almost now a commercialized type of fraud. We call them paper mills: It’s an industry that creates these fake papers and sell the authorships to authors who want to have a paper. They actually advertise on Facebook or other social media. And I think it's a way that folks in maybe countries that do not have a lot of money to give to science, they're hoping for maybe, you know, a better quality of life, maybe a position in North America or in Europe, and they need those papers to pad their resume.

Ryssdal: Explain, then, the concept for us of peer review, because I had thought the whole thing was that people with expertise in the subject at hand review these things, and then they get published.

Bik: No, yes, that is definitely the case. But I think peer review is not really designed to catch fraud. It's designed from the sort of the idea that the data that you're reviewing as a peer reviewer is real. And I think that worked really well until maybe one or two decades ago because the field of science — and specifically the field you're peer reviewing in, which is your own field — was relatively small. But nowadays, there's so many more papers being published. And I think peer reviewers don't realize that there's a fraction of the papers that they might peer review are fraudulent.

Ryssdal: Yeah, a fraction. But still, it's the credibility thing that gets affected here.

Bik: It does, and talking about science misconduct could lead people to think that all science is flawed, but that's not the case. I do want to stress that, and it's really important to realize we're talking about a fraction; a couple of rotten apples in a fruit basket. But I like to think that the rest of the fruit basket is still beautiful. But, yeah, part of it is bad.

Ryssdal: And about the literature, there is a financial incentive for professional journals to publish papers, right? I mean, it's literally part of their business model. They’ve got to do it at scale.

Bik: Yes. And the whole scientific publishing industry is very strange if you explain that to a nonscientist. Because we scientists, we write the papers for free and we peer review them for free. We might even —

Ryssdal: Wait, wait, sorry, peer reviewers don't get paid?

Bik: Yes, usually not. No. I mean, there's some journals that start doing it. But basically, we provide and write and peer review the papers for free, and then we have to pay the scientific publishers to be able to read them or make them open access so that everybody can read them. The scientific publishing industry makes huge amounts of profits. So in return, you would hope what the publishers add to the paper is maybe some quality control or some, you know, fraud detection. And it seems that, yeah, they're not doing a very good job there.

Ryssdal: Do you imagine, then, Dr. Bik, that this problem gets better or worse? Will AI, as it's coming, help you ferret out these frauds? Or will it actually make the problem more challenging for you?

Bik: Both, unfortunately. So we are using tools, AI-based tools, to help find fraudulent papers. But as you can imagine, it's very easy now using AI to create a fake image, and you can make a peer reviewer believe that the experiments that have been done are real. And that is, we assume, already a big problem, but it's so hard to recognize those images.

Ryssdal: Do you imagine you're going to research and integrity consult your way out of a job someday? That'd be your ideal, wouldn't it?

Bik: That would be great if, you know, if there was an actual quality control and we didn't have to do this type of work. And as I'm sort of close to retirement, yeah, I hope that they will come soon. But I think there will be enough work for the next generation of sleuths.