The Impact of a Ghostwritten Paper on the Fate of Glyphosate

By Alexander Kaurov and Naomi Oreskes. Kaurov is a Ph.D. astrophysicist and Ph.D. candidate in the School of Science in Society at Victoria University of Wellington, specializing in computational analysis of scientific discourse. Oreskes is a historian of science at Harvard University whose scholarship traces corporate influence on research and regulation, and author of the books “Merchants of Doubt” and “Why Trust Science?” Originally published at Undark. 

In October 2026, the Environmental Protection Agency must release its decision on the use of America’s most widely used herbicide, glyphosate. It will mark a milestone in the 15-year registration review cycle for pesticides (the umbrella legal term in the U.S. which includes herbicides, insecticides, fungicides, and rodenticides) mandated under federal law. The deadline for the decision, originally scheduled for 2022, was extended to 2026 after the 9th U.S. Circuit Court of Appeals ordered the EPA to reconsider its preliminary conclusion that glyphosate was “not likely” to cause cancer.

This time, however, the EPA heads into the review under dramatically reduced capacity. President Donald Trump’s 2026 budget proposes a 55 percent cut for the agency. In July 2025, the Trump administration began to dismantle the agency’s Office of Research and Development, with plans to lay off more than 3,700 employees — roughly three‑quarters of its research staff and about a fifth of its total workforce. Former EPA administrators warn that this will strip the agency of its in‑house toxicologists, chemists, and epidemiologists — the experts who generate much of the primary data that undergird almost every rule the agency writes.

Besides increasing the possibility that glyphosate and reviews of other pesticides will be further delayed, what else can we expect from this situation? Already, we’ve seen numerous reversals of policy and cancellations of data collection projects under the new EPA administrator, Lee Zeldin. Therefore, it is reasonable to expect more antiregulatory moves.

Limiting the agency’s internal scientific capacity increases its reliance on external expertise and scientific literature, so the agency will have to trust the robustness of published research. But is the scientific record robust enough?

Peer review is supposed to safeguard the accuracy of published science, including keeping it clean from contamination by paper mills, undisclosed conflicts of interest, manipulated data, corporate misconduct, and other forms of malpractice. Unfortunately, the scientific literature has proven far too easy to compromise.

Consider a single review paper published in Regulatory Toxicology and Pharmacology in 2000 about Roundup, the trade name for Monsanto’s widely used glyphosate-based herbicide. Authored by three researchers — Gary M. Williams, Robert Kroes, and Ian C. Munro — who disclosed no conflicts of interest, the paper concluded that “under present and expected conditions of use, Roundup herbicide does not pose a health risk to humans.”

In 2017, internal corporate emails released during federal litigation against Monsanto revealed that the paper was largely conceived and drafted by Monsanto employees. (The company denies this, but the evidence is overwhelming.) In these communications, employees praised one another for contributing to the manuscript, including “writing” it — despite not being listed as authors and only thanked in the acknowledgements for providing “scientific support.”

In short, the paper was ghostwritten — a clear violation of any imaginable standard of scientific ethics.

A Monsanto employee expressed hope that the review would become “‘the’ reference on Roundup and glyphosate safety,” and it did. In our recent research published in Environmental Science and Policy, we show that this paper is in the top 0.1 percent of cited academic literature on glyphosate. The vast majority of papers that cite it offer no acknowledgment of its questionable origins. This fraudulent paper has become deeply integrated into and influential in the scientific record.

The paper’s influence has spread far beyond academia. Government documents from public health agencies around the world — including the Centers for Disease Control and Prevention and Health New Zealand: Te Whatu Ora — cite this ghostwritten paper without caveats even after the 2017 revelations, affecting policy and shaping public perception of glyphosate’s safety. A 2011 Canadian Forest Service publication, for instance, answers a question about whether glyphosate causes cancer and is an endocrine disruptor with the following:

No. Based on the weight of available scientific evidence, several regulatory and independent scientific review panels conclude that glyphosate is non-carcinogenic, does not cause birth defects or genetic alterations, and does not act as an endocrine disruptor in whole animal systems under realistic exposure regimes.

The answer references the paper and continues, saying “Such reviews conducted by highly qualified professional toxicologists and risk assessment specialists provide the most credible and reliable sources of information.”

The public has also been influenced by this paper. Who among us hasn’t turned to Wikipedia for information? It is a frequent top Google search result, and now it is part of the training datasets for many artificial intelligence models. The paper is mentioned in popular Wikipedia articles about Roundup and glyphosate-based herbicides (though there are ongoing attempts to remove it). We analyzed the editing history of these entries and found that although several editors had attempted to note the review’s ghostwritten origins, these notes were systematically excluded by higher level editors.

On Wikipedia’s discussion pages, users report how influential these Wikipedia articles are in their local communities. One comment reads: “The content of this article is dangerous. I work in the agricultural sector in Southern France. I was at a meeting with some farmers discussing safety when a guy addressed the crowd and literally quoted this article stating that glyphosate does not cause cancer and is less dangerous than table salt.” It goes on to say, “This article is used by active farmers as an excuse not [to] bother with safety equipment and appropriate practices.”

In 2018, after the revelations, some Wikipedia editors expressed frustration when the paper persisted as a reference: “I’m not sure why editors are pushing so hard for inclusion of this particular source when multiple non-controversial and more authoritative sources exist for this content.”

Wikipedia’s editorial guidelines encourage citing peer-reviewed literature, but the rules on which sources should be used are flexible. Its editors justify including the paper on the grounds that it was published in a peer-reviewed journal and has never been retracted.

The journal the paper appeared in had been previously implicated in scandals for publishing industry-friendly studies, specifically ones for the tobacco industry. One 2017 analysis showed that 96 percent of the tobacco or nicotine papers in this journal published between January 2013 and June 2015 had authors with tobacco industry ties, and none of the papers drew negative conclusions. Since then, the journal has changed editors and now states: “Regulatory Toxicology and Pharmacology, as the journal serving developments for improvement of human health and environment, will not consider manuscripts that have been supported by tobacco companies.”

But what about papers supported by other companies with an interest in promoting their products, even in the face of science that shows their harms? What about companies that are trying deliberately to manipulate science and regulatory decision-making?

We have formally submitted a retraction request for the Roundup paper to the current editors of the journal, and they have promised to review the case. But this is just one example among what seems to be a growing number of papers contaminating the scientific literature. And it’s doubtful that retractions alone can compensate for the sheer volume of questionable research now circulating.

Glyphosate, moreover, is just one of many herbicides and other pesticides for which the EPA is expected to make regulatory decisions in the near future. Even before the current dismantling of its scientific infrastructure, the EPA’s stance on glyphosate had drawn criticism for being out of step with the World Health Organization’s International Agency for Research on Cancer, which classifies glyphosate as “probably carcinogenic to humans.”

Adding further complexity is the involvement of Secretary of Health and Human Services Robert F. Kennedy Jr., a central figure in the “Make America Healthy Again” movement. He served as co-counsel in a 2017 lawsuit against Monsanto in state court in Alameda County, California, representing plaintiffs seeking damages for non-Hodgkin’s lymphoma allegedly caused by exposure to Roundup. Regulation of agricultural chemicals also figured prominently in his presidential campaign. Some of his recent statements, however, raise questions about whether he still intends to prioritize the issue.

In the current environment, options for recourse may be limited. The public comment period on glyphosate is expected to open later this year or early next year. During the last round of consultation in 2019, the EPA received 283,300 comments across 12,000 individual submissions. Some studies suggest that public attention during these periods can influence regulatory decisions. Leaving a substantive comment can matter. But how is the public expected to write one when any attempt to “do your own research” leads straight into a compromised literature, whether through search or AI conversation, both of which heavily rely on Wikipedia and the academic corpus?

In the long run, the scientific community must step up to protect the integrity of science as an independent and objective enterprise. With governmental scientific capacity decimated, individual researchers, scientific unions, and professional associations must take a stronger stand to ensure that scientific literature remains a reliable foundation for critical decisions.