How the Proliferation of Fraudulent Scientific Papers Is Threatening the Integrity of Cancer Research

There is a perception among many scientists that scientific fraud is a rare occurrence, resulting from the actions of a few isolated bad actors. However, an extensive investigation by Reese A.K. Richardson, PhD, postdoctoral fellow at the Center for Science of Science and Innovation, Kellogg School of Management at Northwestern University, and his colleagues found that large-scale, systematic fraud is happening on an industrial-sized level. The growing prevalence of paper mills (for-profit entities that mass produce and sell fake or manipulated research papers to academics), brokers (go-betweens involved in publishing bogus papers), and predatory journals that do not implement any quality controls on submissions are turning scientific fraud into a lucrative business. According to Dr. Richardson’s research, the number of fraudulent scientific papers appears to be doubling every 1.5 years, faster than the increase in scientific papers overall, which is doubling every 15 years.¹

Cancer is probably the most vulnerable field for fraudulent research.

— REESE A.K. RICHARDSON, PhD

Tweet this quote

To put that percentage in perspective, according to a study published in Nature, at least 400,000 papers published between 2000 and 2022 show the telltale signs of having been produced by paper mills,² yet only 55,000 were retracted or corrected during that same time period.³ And although some research papers that are retracted may be the result of innocent errors rather than intentional misconduct, the overwhelming majority, over 80%, are the result of fraud or plagiarism, according to Dr. Richardson. In addition, the topic of cancer is a prime subject for misinformation research, given its high rate of prevalence worldwide and impact on society.

“Cancer is probably the most vulnerable field for fraudulent research,” said Dr. Richardson. “And, to that extent, a huge fraction of the cancer literature is probably completely unreliable. We have observed a lot of paper mill activity in cancer research, particularly in noncoding RNAs and their role in cancers. Also, there are hundreds of types of cancers and thousands of different drug candidates and compounds to treat cancers, making it easy to choose a combination, fabricate or plagiarize data and images, and then write a plausible-sounding manuscript based on that combination.”

Ensuring Accuracy in ASCO Publications

To help ensure that all manuscripts submitted to ASCO’s scientific journals contain accurate, rigorously researched studies, in 2023, the Journal of Clinical Oncology (JCO) published guidance on the appropriate use of large language models and artificial intelligence (AI) tools in works submitted.⁴ ASCO has also published an Author Conduct Policy to guarantee that all journal submissions meet legal requirements and accepted ethical and scientific standards.

Understanding the Scope of Bogus Research

In a wide-ranging interview with Dr. Richardson, he discussed the rapid increase in fraudulent scientific papers, the incentives for scientists to produce bogus research, and how to spot and combat problematic research and scientific misconduct.

Combating the Industry of Scientific Fraud

Please talk about the advent of paper mills and brokerages or go-betweens involved in publishing bogus or fraudulent scientific papers. What’s driving this business? Is it the pressure on scientists to “publish or perish”; the rapid growth of science; or the availability of new technology like AI, which can seamlessly create or doctor images and texts?

It’s all of the above. Any high school or college student is probably familiar with online services that offer to write your essay, term paper, or even your PhD thesis. It’s largely believed that the industry of fraud in scientific publishing sprang up from this practice.

There has always been fraud in scientific papers, but now the practice is feared to be reaching epidemic proportions, in part because of the pressure on scientists to continuously publish their work. Paper mills realized that they could produce studies that are 100% fraudulent and charge a fee for people to put their name on the paper. These same paper mills will sell citations, thereby increasing the client’s perceived scientific impact.

There has always been fraud in scientific papers, but now the practice is feared to be reaching epidemic proportions, in part because of the pressure on scientists to continuously publish their work.

— REESE A.K. RICHARDSON, PhD

Tweet this quote

The availability of new technologies, including generative AI, certainly lowers the barrier to entry into the industry of scientific fraud. They also make it easier for individuals to manufacture an entire fraudulent study on their own and more difficult to detect for publication gatekeepers.

That said, the trends we studied happened almost entirely before the advent of generative AI, starting in late 2022. Paper mills were an expanding problem well before these technologies entered the publishing scene, so there’s not a complete sea change now that generative AI is much more accessible. However, as these technologies become more sophisticated, the problem is only going to get worse.

There are still big unknowns about this industry, including how long it has been around and its full extent in the publication of fraudulent scientific papers. Generally, the emergence of paper mills first started being documented around 15 years ago.

Feeling Pressure to Publish or Perish

Why would a scientist agree to publish a fraudulent paper? What’s happening in the field of science that is making this practice relatively commonplace?

Many scientists around the world are in situations in which they are being pressured to publish more studies, secure more grants and other research funding, and collaborate internationally. However, they don’t have the resources needed to produce genuine science.

We need collective action from all the stakeholders to find new ways to assess and reward researchers to remove the professional and monetary incentives at the heart of industrialized scientific misconduct.

— REESE A.K. RICHARDSON, PhD

Tweet this quote

A classic example of a paper mill client is a physician in a tertiary hospital in China, where basically he or she is expected to publish in addition to practicing medicine. Most of these clinicians will not have a research laboratory, nor a research staff. They won’t have time to raise funding to do research, but their employer is telling them they need to perform research studies and publish their work. Until recently, medical institutions in China frequently offered cash bonuses to scientists who published a study in a particular journal of a notable caliber. That practice is now banned in scientific research in China, but it’s still very common in other parts of the world, like India.

Coming to America

Is the publication of fraudulent scientific research a big problem in the United States?

Generally, scientists in the United States have the resources to perform genuine scientific research, so they are not under as much pressure as scientists in other countries. However, scientists here are starting to have to compete for an increasingly scarce pool of resources, especially with the funding cuts underway from the current administration, so the pressure is only going to get worse. Anything that increases the amount of competition in the scientific enterprise is going to lead to problems like we’re seeing with the proliferation of fraudulent scientific papers.

Spotting the Red Flags of Bogus Research

A study, in 2023, found that the number of scientific papers published reached nearly 3 million.⁵ What are the telltale signs of large-scale fraud in these papers? What red flags and patterns should scientific journals look for?

Journal editors should look out for sudden increases in the volume of submissions, especially if these submissions all seem to come from the same country and are highly similar. There are also specific red flags that editors can look out for, such as “tortured phrases”: nonsensical, unnatural-sounding word combinations that are often generated by AI to bypass plagiarism detection software looking for verbatim matches. For example, “artificial intelligence” might become “counterfeit consciousness.” Journal editors and publishers should also look for problematic references and reagents, implausible coauthorships, conflicts of interest between editors and peer reviewers, unusually quick publication timelines, and duplicated images.

I recommend that anyone who regularly engages with scientific literature use PubPeer (https://pubpeer.com) and its browser extension for flagging potentially problematic papers. My colleagues and I recently launched the Collection of Open Science Integrity Guides (COSIG; cosig.net), which covers information on many of these red flags.

I don’t want to give the impression that fraudulent scientific papers are only coming from outside of the United States. We have seen instances of prominent cancer researchers in the United States who have had their papers retracted because of intentional misinformation. Although paper mills will probably find the most clients in places where the pressure to publish is greatest and the resources for genuine science are the scarcest, there is no geographic location that is off limits to the paper mill industry.

Guarding Against the Publication of Fake Research

Among the journals you report on in your study is PLOS One, in which you found that 33 editors were responsible for work that was later retracted or criticized on PubPeer. In one instance, of the 79 papers handled by one editor, 49 have been retracted.¹ Now that studies are showing an epidemic of fraudulent research being published, are steps being taken industry-wide to prevent the intentional publication of bogus research?

There are steps being taken to detect and eliminate the possibility of publishing fake research. PLOS One is actually one of the publications that has taken proactive steps to guard against these false studies. Most of the editors we identified were removed from the PLOS One editorial board even before our study took place.

PLOS One had identified some of these patterns beforehand and had taken steps to mitigate the problem. The PLOS team is quite dogged in weeding out fraudulent papers and very responsive to publication integrity concerns. However, that is not the case at other publishers. Some publishers seem not to respond to publication integrity concerns at all. There is no uniform approach that publishers have taken to identify bogus papers, and publishers don’t have a shared commitment to the integrity of their catalog.

Helping Patients Discern Between Legitimate and False Cancer Information

A recent study found that cancer misinformation is rampant on social media sites, including Facebook, Reddit, Pinterest, and Twitter (now X), and that patients with cancer are increasingly turning to AI chatbots like ChatGPT for medical information.⁶ What can the oncology community do to help their patients evaluate the accuracy of the information they find online?

This is a really difficult question to answer. Everyone should be aware of the widespread problem of false research being published both online and in print. For oncology in particular, the problem is especially challenging because there is a lot of fraudulent research that prop up specific supplements and compounds such as cinnamon or curcumin as alternative treatments for cancer that patients will see and believe. Patients assume because they read about these treatments on a website that is quoting information from a “peer-reviewed study” asserting that these compounds are effective, the information is true. The problem will only get worse now that budget cuts, the cancellation of research grants, and massive layoffs underway at the federal health agencies threaten to compromise once trustworthy government resources. Oncologists can help their patients evaluate reputable cancer resources by having open patient-centered conversations and directing them to trustworthy sites, such as the American Cancer Society (cancer.org).

Protecting the Integrity of Science

Please talk about how the publication of fraudulent research is contributing to the public’s growing mistrust of science.

We need to develop measures to protect the integrity of the scientific literature. The first step is to acknowledge that there is a problem. We need the active engagement of the scientific community to stand up for science and to be able to defend it from threats not only coming from federal health agencies, but perpetrated by the thousands of fraudulent studies proliferating the scientific literature. We need collective action from all the stakeholders to find new ways to assess and reward researchers to remove the professional and monetary incentives at the heart of industrialized scientific misconduct.

DISCLOSURE: Dr. Richardson reported no conflicts of interest.

REFERENCES

1. Richardson RAK, Hong SS, Byrne JA, et al: The entities enabling scientific fraud at scale are large, resilient, and growing rapidly. Proc Natl Acad Sci U S A 122:e2420092122, 2025.
2. Van Noorden R: How big is science’s fake-paper problem? Nature, November 6, 2023. Available at www.nature.com/articles/d41586-023-03464-x. Accessed October 15, 2025
3. Abalkina A, Aquarius R, Bik E, et al: ‘Stamp out paper mills’—Science sleuths on how to fight fake research. Nature 637:1047-1050, 2025.
4. Miller K, Gunn E, Cochran A, et al: Use of large language models and artificial intelligence tools in works submitted to Journal of Clinical Oncology. J Clin Oncol 41:3480-3481, 2023.
5. Hanson MA, Barreiro PG, Crosetto P, et al: The strain on scientific publishing. Quantitative Science Studies 5:823-843, 2024.
6. Farrar O: The ‘Wild West’ of cancer misinformation online. Harvard Magazine, February 10, 2025. Available at www.harvardmagazine.com/2025/02/harvard-online-cancer-misinformation#:~:text=The%20findings%2C%20he%20says%2C%20indicate,of%20the%20content%20was%20dangerous. Accessed October 15, 2025.

ASCO’s Guidelines to Ensure Research Integrity in Manuscript Submissions to JCOJournals

In response to the growing use of large language models and artificial intelligence (AI) tools in the creation of research papers, in 2023, the Journal of Clinical Oncology (JCO) published specific guidance on the appropriate use of these devices for manuscripts submitted to JCO.¹ The guidance includes the following statements:

• Authors remain responsible for the accuracy of all content submitted and are liable for any breach of publication ethics;
• JCO does not accept manuscripts with nonhuman authors, and large language models and AI tools cannot be listed as an author under any circumstances;
• JCO generally discourages the use of large language models and AI tools to generate written content in submissions;
• JCO forbids the use of large language models and AI tools in the preparation of submissions primarily advancing the author’s opinion and perspective, such as Art of Oncology and Comments and Controversies;
• Reviewers may not use large language models and AI tools when reviewing work submitted to JCO for peer review.

In addition, to ensure the integrity of research published in all ASCO journals, including JCO, JCO Oncology Practice, JCO Global OncologyJCO Precision Oncology, JCO Clinical Cancer Informatics, and JCO Oncology Advances (“the Journals”), and safeguard against the introduction of fraudulent research, ASCO has developed the Author Conduct Policy (https://ascopubs.org/pb-assets/pdfs/author-conduct-policy-1545422401880.pdf). The Author Conduct Policy includes detailed information on author responsibilities for manuscript submissions, including adherence to ASCO’s Conflict of Interest Policy (www.asco.org/about-asco/legal/conflict-interest-coi). In addition, the Author Conduct Policy outlines ASCO’s oversight system to determine potential author misconduct, as well as an approach to address author concerns based on guidelines established by the International Committee of Medical Journal Editors (ICMJE).

Author misconduct in manuscript submissions is defined as:

• Fabrication or falsification of data;
• Plagiarism, either of one’s own work, ideas, processes, results, or words or those of another person;
• Redundant publication of content that overlaps substantially with content already published in print or electronic media;
• Failure to credit an author, including removal of an author without the author’s consent and the consent of other coauthors;
• Inappropriate credit of an author, including adding as a “gift” or “guest” author an individual whose contribution does not meet the Journals’ definition of authorship, as established by the ICMJE;
• Submitting as one’s own work “ghostwritten” manuscripts, in which the listed author(s) is unfamiliar with the underlying data and relevant documentation and had minimal or limited input in the development and/or writing of the article;
• Treatment of research subjects in a manner that does not comport with applicable laws or accepted standards of ethics;
• Failure to disclose relationships in accordance with ASCO’s Conflict of Interest Policy.

“ASCO strives continuously to be a trusted source of oncology information,” said Clifford A. Hudis, MD, FASCO, FACP, Chief Executive Officer of ASCO and Executive Vice Chair of the Conquer Cancer

Clifford A. Hudis, MD, FASCO, FACP

Foundation. “For JCO Journals, that means our editorial teams follow a robust methodology to ensure the quality and importance of its articles. The approach includes a rigorous peer-review process and scanning accepted articles for plagiarism. JCO Journals’ editors have a clear and unwavering policy on author conduct that includes applying strong sanctions as appropriate, such as restricting authors from publishing.”

Dr. Hudis continued: “Trust in the integrity of scientific research is a serious issue. Clinicians, practitioners—and ultimately patients—count on the credibility of that research to inform cancer care guidelines and practices, and we work to earn that trust every day.”

DISCLOSURE: Dr. Hudis reported no conflicts of interest.

REFERENCE

1. Miller K, Gunn E, Cochran A, et al: Use of large language models and artificial intelligence tools in works submitted to Journal of Clinical Oncology. J Clin Oncol 41:3480-3481, 2023.