Crimes Against Science

Introduction: what counts as a crime against science

Science is a trust system. A reader cannot rerun every experiment, so each published claim borrows credibility from the honesty of whoever made it. A crime against science is any act that spends that credit dishonestly: inventing data, bending it, stealing it, extracting it from people who never consented, or inflating the signals by which science judges its own work.

The modern definition used by the United States Office of Research Integrity (ORI) names three core offences: fabrication (making up data), falsification (manipulating materials, equipment, processes or results so the record misrepresents the research) and plagiarism (taking another person's ideas, processes or words without credit). This narrow triad, often shortened to FFP, is a useful legal floor but not the whole picture. History shows at least three further families of violation: the abuse of human beings as research material, the capture of science by ideology or the state, and the corruption of the metrics that decide whose work counts.

This essay follows those violations roughly in chronological order. It begins with disputed data in antiquity and early modern science, passes through hoaxes, ideological science and atrocities against research subjects, examines the industrial-scale fraud of recent decades, and ends with the quietest offence of all: the manipulation of citations, where no single number is false but the whole ledger of scientific reputation is rigged.

One caution runs throughout. Historical judgement is easy to overplay. Some celebrated "frauds" are better read as the normal practices of their era, or as honest error. Where a case is genuinely contested, the essay says so.

Part I. Early violations: from antiquity to the Enlightenment

The oldest accusations of scientific dishonesty concern data that fit theory too neatly. They are also the hardest to judge, because the norms of reporting observations were different, and the accused cannot answer.

Ptolemy and the too-perfect sky

Claudius Ptolemy's Almagest (2nd century CE) dominated astronomy for fourteen centuries. In The Crime of Claudius Ptolemy (1977), the physicist Robert R. Newton argued that many of Ptolemy's "observations" were computed backwards from his theory and partly borrowed from Hipparchus. Other historians reply that ancient astronomers presented illustrative, worked values rather than raw logbooks. The case remains open, but it established the template for every later suspicion: results that agree with the model better than real measurement allows.

Galileo and the suppression of evidence

Not every early crime was committed by a scientist. In 1633 the Roman Inquisition forced Galileo Galilei to abjure heliocentrism and placed him under house arrest, and his Dialogue went onto the Index of Prohibited Books. This is the archetype of the external crime against science: an authority deciding in advance which conclusions evidence may reach. It recurs, with different authorities, throughout this essay.

Beringer's lying stones (1725)

Johann Bartholomeus Adam Beringer, a physician and professor at the University of Würzburg, published a lavish volume on carved "fossils" showing lizards, stars, and Hebrew letters. They had been carved and planted by two colleagues, J. Ignatz Roderick and Johann Georg von Eckhart, reportedly to humiliate him. Beringer was the victim rather than the perpetrator, but the episode shows how a scholar's desire for spectacular evidence can lower his guard. The case went to court, and the reputations of all three men suffered.

Newton: the priority war and the fudge factor

Isaac Newton offers two lessons. First, in the calculus priority dispute with Gottfried Wilhelm Leibniz, the Royal Society appointed a committee in 1712 whose report, the Commercium Epistolicum, found for Newton. Newton was President of the Society and largely drafted the report himself, then reviewed it anonymously. It is an early case of a judge sitting on his own case.

Second, the historian Richard Westfall showed in "Newton and the Fudge Factor" (Science, 1973) that Newton adjusted calculations in later editions of the Principia, for the velocity of sound and the precession of the equinoxes, to claim precision his data could not support. The physics was right; the advertised accuracy was not.

Mendel's peas

In 1936 the statistician Ronald A. Fisher concluded that Gregor Mendel's published ratios were statistically "too good" to be true. Explanations range from unconscious bias in classifying peas, to an assistant who knew the expected answer, to selective reporting of the most typical experiments. Most modern scholars reject deliberate fraud. The case is valuable precisely because it shows that the line between fraud, bias and convention is often a matter of evidence we no longer have.

Babbage names the offences (1830)

The first systematic taxonomy came from the mathematician Charles Babbage in Reflections on the Decline of Science in England (1830). He distinguished four practices:

• Hoaxing: a deliberate deception meant to be discovered, like Beringer's stones.
• Forging: recording observations that were never made.
• Trimming: clipping outlying readings to make results look more precise.
• Cooking: selecting only the data that agree with the hypothesis.

Babbage's vocabulary is still useful. Forging is today's fabrication; trimming and cooking are falsification. Nearly two centuries later, the categories have barely changed.

Part II. Hoaxes, self-deception and forged evidence (1900–1980)

As science professionalised, the motives for deception sharpened: national pride, the defence of a favoured theory, and personal fame. Four cases from this period mark the spectrum from self-deception to deliberate forgery.

N-rays: when a laboratory fools itself (1903)

The French physicist René Blondlot announced a new form of radiation, "N-rays", and dozens of French researchers soon confirmed it. In 1904 the American physicist Robert W. Wood visited Blondlot's laboratory and secretly removed the prism essential to the apparatus. Blondlot continued to "see" the rays. No forgery occurred, but the episode became the classic example of what the chemist Irving Langmuir later called "pathological science": effects at the edge of perception, defended by observers who want them to exist.

Piltdown Man (1912–1953)

Charles Dawson, an amateur collector, presented skull fragments and a jaw from a gravel pit at Piltdown, Sussex. Backed by the British Museum's Arthur Smith Woodward, "Piltdown Man" was hailed as the missing link, conveniently English, with a large brain and an ape-like jaw. For four decades it distorted human evolution, partly because it confirmed what British scientists expected.

In 1953 Kenneth Oakley, Joseph Weiner and Wilfrid Le Gros Clark exposed it using fluorine dating and close inspection. The cranium was medieval human, the jaw belonged to an orangutan, and the teeth had been filed and stained. A 2016 study by Isabelle De Groote and colleagues, using deoxyribonucleic acid (DNA) analysis and imaging, concluded that the specimens came from a single orangutan and pointed to one forger, most plausibly Dawson himself. Piltdown teaches that fraud survives longest when it flatters the community that examines it.

Kammerer's midwife toad (1926)

The Austrian biologist Paul Kammerer claimed that midwife toads (Alytes obstetricans) forced to breed in water developed dark "nuptial pads" and passed them to offspring, evidence for the inheritance of acquired characteristics. In 1926 the herpetologist G. Kingsley Noble examined the one surviving specimen and reported in Nature that the pads had been made with injected India ink. Kammerer died by suicide weeks later. Whether he, an assistant or a saboteur injected the ink has never been settled, and some modern epigeneticists argue his observations deserve a second look. The case still shows how a single doctored specimen can end a career and a research programme.

Cyril Burt and the twins (1950s–1970s)

Sir Cyril Burt, a founder of British educational psychology, published studies of identical twins reared apart that strongly supported the heritability of intelligence. They influenced the selective "eleven-plus" school system. After his death, the psychologist Leon Kamin (1974) noted that Burt's correlation coefficients stayed identical to three decimal places, around 0.771, even as his sample size grew. That is statistically implausible. In 1976 the journalist Oliver Gillie reported that two of Burt's co-authors, Margaret Howard and J. Conway, could not be traced.

Burt's official biographer, Leslie Hearnshaw, concluded in 1979 that Burt had fabricated data. Later defenders, including Robert Joynson and Ronald Fletcher, argued for carelessness rather than fraud. The scientific verdict is that his twin data cannot be trusted, whatever his intent. Burt matters because his numbers shaped the lives of millions of schoolchildren: fraud in social science is never merely academic.

Part III. When the state writes the conclusions

The twentieth century showed that the gravest damage to science comes not from a lone forger but from a regime that decides the answer in advance and punishes those who disagree. Here the crime is institutional, and its victims include whole populations.

Lysenkoism in the Soviet Union

Trofim Lysenko, an agronomist with little formal training, rose under Stalin by promising rapid harvest gains through "vernalisation" and by rejecting Mendelian genetics as bourgeois. His claims rested on anecdote, poorly controlled trials and political endorsement rather than evidence. Opponents paid with their lives. Nikolai Vavilov, the country's leading plant geneticist and a founder of seed-bank science, was arrested in 1940 and died in a Saratov prison in 1943.

At the August 1948 session of the Lenin All-Union Academy of Agricultural Sciences, Lysenko announced that the Central Committee had approved his report. Genetics was effectively outlawed, geneticists lost their posts, and textbooks were rewritten. Soviet biology did not recover its standing until Lysenko's fall in the mid-1960s. Lysenkoist methods such as close planting were also adopted during China's Great Leap Forward, where they contributed to agricultural collapse. Lysenkoism is the purest case of falsification by decree.

Science in the service of race: Nazi Germany

The Nazi state corrupted science in three ways. First, it purged it: the 1933 Law for the Restoration of the Professional Civil Service removed Jewish and politically suspect academics, driving out figures such as Albert Einstein, Fritz Haber and Lise Meitner, and hollowing out German physics, chemistry and medicine.

Second, it promoted ideological pseudo-science. The Nobel laureates Philipp Lenard and Johannes Stark championed Deutsche Physik, which denounced relativity and quantum theory as "Jewish physics". Racial hygiene, built on an international eugenics movement that also flourished in the United States and Scandinavia, was elevated into state doctrine and used to justify compulsory sterilisation of roughly 400,000 people.

Third, institutions lent scientific authority to murder. The Kaiser Wilhelm Institute for Anthropology, Human Heredity and Eugenics, under Eugen Fischer and later Otmar von Verschuer, supplied the theoretical frame; von Verschuer received specimens from his former assistant Josef Mengele at Auschwitz. The "T4" programme killed tens of thousands of disabled patients, with physicians selecting the victims. This leads directly to the darkest chapter: crimes committed against human beings in the name of research.

Part IV. Crimes against the human subject

Fabrication corrupts the record; human experimentation without consent corrupts the purpose of science itself. These cases are crimes in the ordinary legal sense, and they produced the ethical architecture that governs research today.

The Nazi camp experiments and the Doctors' Trial

In the camps, physicians used prisoners as experimental material. At Dachau, Sigmund Rascher ran high-altitude experiments in a low-pressure chamber and freezing experiments in ice water for the Luftwaffe, killing many subjects. Others conducted sulfonamide and bone-grafting experiments at Ravensbrück, malaria and mustard-gas studies, and Mengele's twin research at Auschwitz.

At the Doctors' Trial in Nuremberg (1946–1947), 23 defendants, 20 of them physicians, were tried; 16 were convicted and seven were hanged. The judgment set out the Nuremberg Code (1947), whose first principle is that the voluntary consent of the human subject is absolutely essential. A lasting ethical question followed: whether data obtained by torture, such as Rascher's hypothermia measurements, may ever be cited. Many journals and ethicists hold that such citation, if it occurs at all, must name its origin.

Unit 731 and the bargain of silence

Imperial Japan's Unit 731, commanded by the army physician Shirō Ishii near Harbin in occupied Manchuria, conducted vivisection, deliberate infection with plague, anthrax and cholera, frostbite experiments and weapons trials on prisoners, mostly Chinese. Thousands died in the facility and many more in field attacks. After the war, the United States granted Ishii and his colleagues immunity from prosecution in exchange for their data. That bargain is itself a crime against science: it treated knowledge extracted by atrocity as an asset to be traded rather than evidence to be judged.

Tuskegee (1932–1972)

The United States Public Health Service (USPHS) followed about 600 Black men in Macon County, Alabama, roughly 400 with latent syphilis and 200 controls, to observe the disease's natural course. The men were told they were being treated for "bad blood". When penicillin became the standard cure in the late 1940s, it was withheld, and the men were steered away from treatment elsewhere. The study ran until the journalist Jean Heller exposed it in 1972.

Tuskegee led to the National Research Act (1974), which established Institutional Review Boards (IRBs) as a federal requirement, and to the Belmont Report (1979), with its principles of respect for persons, beneficence and justice. President Bill Clinton apologised formally in 1997. Its legacy of distrust in medicine among Black Americans persists.

Guatemala (1946–1948)

While Tuskegee was running, the USPHS physician John Cutler led studies in Guatemala that deliberately infected about 1,300 prisoners, soldiers, psychiatric patients and sex workers with syphilis, gonorrhoea and chancroid. The work was hidden until the historian Susan Reverby found Cutler's papers in 2010, prompting a formal apology from the United States government. Cutler later worked on Tuskegee.

Abuse in ordinary hospitals

The violations were not confined to dictatorships or remote programmes. In 1966 the anaesthetist Henry Beecher published "Ethics and Clinical Research" in the New England Journal of Medicine (NEJM), describing 22 unethical studies drawn from mainstream journals and leading institutions. At Willowbrook State School in New York (1956–1971), children with intellectual disabilities were deliberately infected with hepatitis, with parental "consent" sometimes tied to admission. In 1951 cells taken from Henrietta Lacks, a Black woman treated for cervical cancer at Johns Hopkins, became the immortal HeLa line without her knowledge; her family learned decades later.

The World Medical Association's Declaration of Helsinki (1964, revised repeatedly) and Beecher's exposé moved consent from a wartime lesson to a peacetime rule. The core principle these cases established is simple: a scientific result never justifies the means used to obtain it.

Part V. The age of the serial fabricator (1974–present)

From the 1970s, the expansion of grant funding, publication counts and laboratory hierarchies created a new figure: the productive star whose output was largely invented. Detection improved too, so the modern record is better documented than any earlier era.

The painted mouse and the cardiology prodigy

In 1974 William Summerlin of the Sloan-Kettering Institute claimed to have transplanted skin between genetically different mice without rejection. A technician noticed that the dark "grafts" on white mice washed off with alcohol: Summerlin had coloured them with a felt-tip pen. In 1981 John Darsee, a young cardiologist at Harvard, was seen fabricating data; investigations found invented results across years of work at Emory and Harvard, with dozens of papers and abstracts withdrawn. The Darsee case exposed "honorary" co-authors who had signed papers they had not checked, and helped push the United States to create formal misconduct offices, eventually the ORI in 1992.

Schön and the physics of copy-paste (2002)

Jan Hendrik Schön of Bell Laboratories published a stream of breakthroughs in organic semiconductors and superconductivity, at one point roughly a paper every eight days, many in Science and Nature. Outside physicists noticed that the same noise curves appeared in graphs from supposedly different experiments. A 2002 committee chaired by Malcolm Beasley found misconduct in 16 of 24 allegations. Schön had kept no laboratory notebooks and had deleted raw data. Some two dozen papers were retracted, and the University of Konstanz revoked his doctorate. Peer review, it turned out, had checked plausibility, not reality.

Hwang Woo-suk: fraud and coercion together (2004–2006)

The South Korean veterinarian Hwang Woo-suk reported in Science in 2004 and 2005 that he had cloned human embryos and derived patient-specific stem-cell lines. He was a national hero. In 2005–2006 investigations by Seoul National University showed that the stem-cell lines did not exist as claimed and that the data were fabricated. They also found that eggs had been obtained from junior female researchers in his own laboratory, a violation of consent through institutional pressure. Hwang combined the two great crimes of this essay in one case: falsifying the record and abusing research subjects.

Wakefield: the fraud with a body count (1998–2010)

In 1998 Andrew Wakefield and colleagues published a paper in The Lancet describing 12 children and suggesting a link between the measles, mumps and rubella (MMR) vaccine and a new syndrome of bowel disease and autism. Investigations by the journalist Brian Deer revealed undisclosed payments from lawyers preparing litigation against vaccine makers, altered patient histories, and invasive procedures performed without proper ethical approval. The Lancet fully retracted the paper in 2010, and the General Medical Council struck Wakefield off the United Kingdom medical register the same year.

The damage outlived the retraction. Vaccine confidence fell, and measles returned to communities where it had been eliminated. Wakefield shows that a single fraudulent paper can harm people who never read it.

The record holders: anaesthesia's twin disasters

The two most-retracted researchers in history are both anaesthesiologists. The Japanese researcher Yoshitaka Fujii was exposed in 2012, partly through a statistical analysis by the British anaesthetist John Carlisle, who showed that the baseline variables in Fujii's trials were too uniform to be real. The German anaesthesiologist Joachim Boldt, a leading figure in fluid therapy, was found in 2010–2011 to have run trials without ethical approval and fabricated data, including studies of hydroxyethyl starch.

As of July 2026, Retraction Watch counts 240 retractions for Boldt, ahead of Fujii at 172 (Retraction Watch). Boldt's case had clinical weight: once his fraudulent studies were removed, meta-analyses showed that hydroxyethyl starch increased the risk of kidney injury and death in critically ill patients. Fraud had shaped bedside practice.

Social science: Stapel, Gino and the replication era

The Dutch social psychologist Diederik Stapel, of Tilburg University, was exposed in 2011 by junior colleagues. He had invented entire datasets for years, often handing students "data" from studies never run; dozens of his papers were retracted. The case helped trigger psychology's replication crisis and the open-science reforms that followed.

A decade later, the blog Data Colada identified apparent data manipulation in papers co-written by the Harvard Business School professor Francesca Gino, a specialist in honesty. After an internal investigation found misconduct, Harvard revoked her tenure in May 2025, the first such revocation at the university since at least the 1940s (The Harvard Crimson). Gino denies wrongdoing, and parts of her lawsuit against Harvard continued.

Macchiarini: from misconduct to prison

Paolo Macchiarini, a surgeon at Sweden's Karolinska Institute, claimed success with synthetic, stem-cell-seeded tracheas. Patients died, and his papers misrepresented their outcomes. Karolinska dismissed him in 2016 and found him guilty of scientific misconduct in 2018. In June 2023 the Svea Court of Appeal convicted him of gross assault in three cases and sentenced him to two and a half years in prison (AP via KSAT). It is one of the rare cases where research misconduct ended in a criminal conviction for harming patients.

Prestige is no protection

Recent cases have reached the top of the profession. In 2014 the STAP stem-cell papers by Haruko Obokata in Nature were retracted within months. In 2023 Marc Tessier-Lavigne resigned as President of Stanford University after a review found manipulated images in papers from laboratories he led. In 2024 Nature retracted the influential 2006 Alzheimer's paper by Sylvain Lesné and colleagues on the amyloid-β species Aβ*56, after image-manipulation concerns, which had shaped years of research funding. Fraud is not a problem of weak institutions only; it thrives wherever prestige substitutes for checking.

Part VI. Manufactured doubt and the hidden trial

Individual fraudsters invent data. Industries have learned something subtler and often more damaging: how to shape the evidence base without forging a single number, by funding, framing, ghostwriting and burying.

Tobacco: doubt as a product

After studies in the early 1950s linked smoking to lung cancer, tobacco executives met in New York in December 1953 and, in January 1954, published "A Frank Statement to Cigarette Smokers" in hundreds of newspapers. They created the Tobacco Industry Research Committee, presented as independent science, whose real function was to fund research that kept the question "open". A 1969 Brown & Williamson memo stated the strategy plainly: doubt was the product, because it competed with established fact in the public mind. The historians Naomi Oreskes and Erik Conway later showed in Merchants of Doubt (2010) that the same playbook, and sometimes the same scientists, reappeared in debates on acid rain, ozone depletion and climate change.

Sugar and the shifting of blame

In 2016 Cristin Kearns, Laura Schmidt and Stanton Glantz published internal documents in JAMA Internal Medicine showing that the Sugar Research Foundation had paid Harvard nutrition scientists in the 1960s for a 1967 review in the NEJM. The review played down sugar's role in heart disease and emphasised dietary fat. The funding was not disclosed. Nutrition policy for decades leaned on a literature partly steered by those who sold the product.

Vioxx and the missing heart attacks

Merck's anti-inflammatory drug rofecoxib (Vioxx) was supported by the VIGOR trial, published in the NEJM in 2000. Three myocardial infarctions in the Vioxx group were left out of the published analysis, making the cardiovascular risk look smaller. Merck withdrew the drug in 2004 after another trial confirmed the danger, and the NEJM issued an expression of concern in 2005. Litigation documents later showed that Merck staff had drafted many research papers that academic authors then signed, a practice known as ghostwriting. Estimates of excess heart attacks attributable to Vioxx run into the tens of thousands.

Study 329 and the antidepressant record

GlaxoSmithKline's Study 329, published in 2001, reported that paroxetine was safe and effective in adolescents with depression. A 2015 reanalysis of the raw data in the BMJ, under the Restoring Invisible and Abandoned Trials (RIAT) initiative, found it was neither effective nor safe, with suicidal behaviour under-reported. The original paper has never been retracted.

Publication bias: the crime of omission

The largest distortion in medicine may be what is never published. In 2008 Erick Turner and colleagues compared antidepressant trials registered with the United States Food and Drug Administration (FDA) with the published literature, in the NEJM. Almost all positive trials were published; most negative ones were either unpublished or written up as if positive. A reader of the journals would conclude that the drugs worked far better than the full evidence showed.

The response has been structural. Since 2005 the International Committee of Medical Journal Editors (ICMJE) has required trials to be registered before enrolment, and the AllTrials campaign (2013) pressed for all results to be reported. Compliance remains incomplete. Selective publication shows that integrity is not only a matter of what is said, but of what is allowed to stay silent.

Part VII. Industrial fraud: images, paper mills and predatory journals

The twenty-first century turned fraud from a craft into an industry. Where Schön forged data alone, today's fraud is bought and sold: whole papers, authorship slots, peer reviews and, as the final part shows, citations.

The doctored image

Digital figures made manipulation easy: a Western blot band can be duplicated, a microscopy field flipped and reused, a gel spliced. In 2016 the microbiologist Elisabeth Bik and colleagues screened 20,621 papers in 40 biomedical journals and found inappropriately duplicated images in about 3.8% of them. Volunteer sleuths, working through the platform PubPeer, now find such problems faster than journals do. The Tessier-Lavigne and Lesné cases began this way.

Paper mills

A paper mill is a commercial operation that manufactures manuscripts, often templated and with fabricated data, and sells authorship to researchers who need publications for promotion, graduation or bonuses. Mills place papers by exploiting guest-edited special issues and by recruiting or bribing editors. They disguise copied text with automated synonym swaps, producing "tortured phrases" that the researchers Guillaume Cabanac and Cyril Labbé used as fingerprints: "counterfeit consciousness" for artificial intelligence, or "bogus upside" for false positive.

The scale is now systemic. After Wiley acquired the open-access publisher Hindawi for about $298 million in 2021, it uncovered mill activity across its special issues, retracted more than 11,300 papers, closed journals and retired the Hindawi brand by 2024 (The Register; Retraction Watch). In August 2025 Reese Richardson, Luís Amaral and colleagues reported in the Proceedings of the National Academy of Sciences (PNAS) that suspected paper-mill output was doubling roughly every 1.5 years, against about 15 years for science as a whole, far outpacing retractions (Richardson). Generative artificial intelligence (AI) has lowered the cost of producing plausible text and images still further.

Predatory publishing

Predatory journals charge article-processing fees while performing little or no peer review. The librarian Jeffrey Beall catalogued them on "Beall's List" from 2010 until he took it down in 2017 under pressure. Such journals serve the same market as paper mills: researchers judged by publication counts rather than by what they publish. They also corrupt the citation record, since their papers are indexed, cited and counted.

This is the bridge to the last and quietest crime. Once science measures people by numbers, the numbers themselves become targets of fraud.

Part VIII. The citation economy and its counterfeiters

The last crime against science is the hardest to see, because every paper involved may contain true data. What is falsified is not the result but the record of whose results matter. Citations were designed as acknowledgement of intellectual debt. Once they became currency for jobs, grants, rankings and bonuses, they acquired counterfeiters.

How a footnote became money

In 1955 the information scientist Eugene Garfield proposed a citation index in Science to help researchers trace ideas. In the 1970s his Institute for Scientific Information introduced the Journal Impact Factor (JIF), the average citations to a journal's recent papers, as a tool for librarians choosing subscriptions. In 2005 the physicist Jorge Hirsch proposed the h-index: a researcher has index h if h of their papers have each been cited at least h times.

None of these tools was designed to rank individuals, yet institutions adopted them for exactly that. Global university rankings, such as the Academic Ranking of World Universities (the Shanghai ranking), the Times Higher Education (THE) and Quacquarelli Symonds (QS) rankings, weigh citations and "highly cited" researchers heavily. Many countries and universities pay cash bonuses per paper or per citation. The result is the condition Goodhart's law describes, in the anthropologist Marilyn Strathern's phrasing: when a measure becomes a target, it ceases to be a good measure.

A taxonomy of citation manipulation

The individual: self-citation and the private cartel

Some self-citation is normal: researchers build on their own work. The problem is scale and purpose. In 2019 John Ioannidis and colleagues published a standardised database of citation metrics for about 100,000 scientists, which made extreme self-citation visible, and a Nature analysis by Richard Van Noorden and Dalmeet Singh Chawla identified a few hundred researchers for whom most citations came from themselves or co-authors.

The most dramatic recent case involves Juan Manuel Corchado, a computer scientist who became Rector of the University of Salamanca in 2024. From 2023 the Spanish newspaper El País reported that he had instructed staff to insert citations to his work into their papers and that low-quality documents citing him had been uploaded to university repositories, which Google Scholar indexed. An external report requested by Spain's Research Ethics Committee described a publication "factory", and Springer Nature retracted dozens of conference papers involving him (elDiario.es). Corchado denied wrongdoing and attributed the campaign to politics. The case shows how a citation profile can become a stepping stone to institutional power.

The gatekeeper: coercive citation

Coercion is the abuse of editorial or reviewing power. In a 2012 survey published in Science, the economists Allen Wilhite and Eric Fong found that roughly one in five academics in the social sciences and business reported being asked by an editor to add citations to the editor's journal without a clear scholarly reason, with junior authors most likely to comply. In 2019 Elsevier concluded that Kuo-Chen Chou, an editor at the Journal of Theoretical Biology, had repeatedly asked authors to cite his own papers, and removed him. Reviewers can do the same: a request to "cite these relevant works" may be a demand to cite the reviewer.

The journal: self-citation and stacking

Journals have their own incentive, because the JIF determines submissions and prestige. Clarivate, which publishes the JCR, suppresses the impact factors of journals with abnormal citation patterns. In 2013 it suspended several Brazilian journals whose editors had published articles citing each other's journals heavily, an arrangement that boosted their standing in Brazil's national assessment system.

The problem is growing. Clarivate suppressed four impact factors in 2023, 17 in 2024 and 20 in 2025, the 2025 cases including titles from MDPI, Sage, Springer, Taylor & Francis and Wiley. Four of the 2025 suppressions were for citation stacking and the rest for self-citation (Retraction Watch). In the same release, Clarivate for the first time excluded citations to retracted papers from impact-factor calculations.

The market: bought citations and sneaked references

The newest development is the outright sale of citations. In a 2024 study, Hazem Ibrahim, Fengyuan Liu, Yasir Zaki and Talal Rahwan of New York University Abu Dhabi analysed about 1.6 million Google Scholar profiles. Posing as a fictional author, they bought 50 citations from a citation-boosting service and watched them appear (arXiv). Paper mills now sell citations alongside authorship, so a fake paper can both carry its buyer's name and inflate another buyer's count.

Manipulation can also happen invisibly in metadata. Guillaume Cabanac, Cyril Labbé and colleagues showed in 2023 that some publishers had deposited reference lists with indexing services containing citations that did not appear in the papers' text at all. Citation counts based on that metadata were inflated by references no reader would ever see.

The institution: buying the highly cited

Institutions can game the system without any individual faking a citation. In 2011 Science reported that some Saudi universities were paying highly cited researchers abroad to add the Saudi institution as an affiliation, often in return for short annual visits. Because rankings such as Shanghai count Clarivate's Highly Cited Researchers by affiliation, the effect was to transfer ranking points.

In 2023 El País reported on Spanish researchers who had listed Saudi universities as their primary affiliation. The chemist Rafael Luque, one of the most cited scientists in Spain, was suspended without pay for 13 years by the University of Córdoba for working for King Saud University and a Russian university while holding a full-time public contract there (El País via WION). The following year, the number of Saudi-affiliated researchers on Clarivate's list fell from 109 to 76 after Clarivate tightened its screening, and the Shanghai ranking stopped counting secondary affiliations (Pesquisa FAPESP).

This pattern is not confined to one country. It follows the incentive wherever rankings are tied to national prestige and funding, including cash-per-paper schemes that were common in China until Chinese authorities moved against them in 2020. The researchers who sell affiliations and the institutions that buy them are responding to the same flawed scoreboard.

Why citation manipulation is a crime against science

It is tempting to treat citation games as harmless vanity. They are not, for three reasons.

1. They misdirect resources. Hiring, promotion and grant committees that rely on citation counts reward manipulators over honest researchers, and so select for the behaviour.
2. They corrupt the map of knowledge. Citations are how researchers find relevant work and how systematic reviews decide what matters. Padding them buries genuinely important papers and lends authority to weak ones.
3. They finance fraud. The market for citations and authorship is what keeps paper mills profitable. Without metrics to inflate, there would be little to sell.

Citation manipulation thus completes the arc of this essay. Babbage's "cooking" selected data to fit a theory. Modern citation cooking selects references to fit a career.

Part IX. The fight back: sleuths, watchdogs and their books

Fraud does not expose itself. Most of the recent cases in this essay were found by outsiders: a microbiologist with a sharp eye, an anaesthetist with a spreadsheet, a student reporter, a pseudonym on a comment site. Several have been sued. They deserve credit by name.

The sleuths

Elisabeth Bik, whose 2016 image study appears in Part VII, left her biotech job and became a full-time investigator, funded by consulting, talks and Patreon (Yahoo News). She won the 2021 John Maddox Prize for her work on image manipulation, plagiarism and flawed methods (Springer Nature). The work has also cost her. After she raised concerns on PubPeer about Didier Raoult's 2020 hydroxychloroquine paper, a lawyer acting for Raoult and Eric Chabrière filed a complaint against her in Marseille in April 2021, alleging moral harassment, attempted blackmail and attempted extortion. She denies it. She has also faced harassment on Twitter and threats of legal action (Nature; Science).

John Carlisle, the anaesthetist who analysed 168 of Fujii's trials in 2012, did not stop there. In 2017 he applied the same kind of test to 5,087 randomised trials in eight journals, including JAMA and the NEJM, and flagged about 2% as having data unlikely to arise by chance (Retraction Watch). Critics warned that the method could cast suspicion on honest trials. A flag is a reason to look, not a verdict.

Nick Brown and James Heathers built checks out of plain arithmetic. Their GRIM test, proposed in 2016, asks whether a reported mean is even possible for the stated number of participants. A failed check is not proof of fraud, but the authors have to explain it (GRIM test). With Tim van der Zee and Jordan Anaya, Brown and Heathers documented problems in more than 50 studies by the Cornell food researcher Brian Wansink. In September 2018 Cornell found that he had committed misconduct and removed him from research and teaching, and he resigned (Retraction Watch). Brown also translated Stapel's memoir into English, Faking Science, as a free PDF.

The three Data Colada authors, Uri Simonsohn, Leif Nelson and Joe Simmons, took their concerns about four of Gino's papers to Harvard. Gino then sued Harvard and all three of them for defamation. In September 2024 the court dismissed her claims against the three (Science). In July 2025 it declined to make her pay their legal fees (The Harvard Crimson). They won, and they still paid for the fight. Her contract claim against Harvard continues.

Brian Deer, whose reporting brought down Wakefield's paper, was sued by Wakefield in Texas in January 2012, together with the BMJ and Fiona Godlee. The suit was dismissed that August for lack of jurisdiction (Courthouse News).

In January 2024 Sholto David, an independent image sleuth in Wales, flagged suspect images in more than 30 papers by four senior scientists at Dana-Farber Cancer Institute, including its chief executive. Within three weeks the institute said it was seeking six retractions and 31 corrections. It announced no finding of misconduct. Some of the images had been raised on PubPeer a decade earlier (STAT).

Jennifer Byrne built Seek & Blastn, a tool that checks whether the gene sequences a paper uses match what the paper says they target (PLoS ONE). A 2021 preprint applying it to 11,799 papers found 712, in 78 journals, with at least one wrongly identified sequence (bioRxiv). Anna Abalkina of the Free University of Berlin traced a single Russia-based paper mill: 1,063 co-authorship offers between 2019 and mid-2022, at least 434 papers linked to it, buyers from at least 39 countries (Times Higher Education).

Some work under pseudonyms, such as Cheshire and Inconstans. "Smut Clyde", who flagged hundreds of articles that may come from paper mills, turned out to be David Bimler, a retired perceptual psychologist in New Zealand (Undark). So do journalists. Theo Baker, a student reporter at The Stanford Daily, reported from November 2022 on image problems in papers from Marc Tessier-Lavigne's laboratories, and won a 2023 George Polk Award, its youngest recipient (Wikipedia). In September 2026 Tessier-Lavigne released emails disputing parts of that reporting (The Stanford Daily).

Most of this work is unpaid, and the legal risk falls on the person who speaks first.

The watchdogs and their tools

Ivan Oransky and Adam Marcus started Retraction Watch on 3 August 2010; their nonprofit, the Center for Scientific Integrity, runs it (Retraction Watch). A $400,000 grant from the MacArthur Foundation paid for its database of retractions (CTSNet). In September 2023 Crossref bought the database for an initial $175,000 and an annual fee, and made it free and open through its public interface. It held about 50,000 retractions, and Retraction Watch still updates it (Crossref). Abalkina built its Hijacked Journal Checker in 2022, which lists more than 400 fake sites posing as real journals (Retraction Watch).

PubPeer, launched in October 2012, lets anyone comment on a published paper, anonymously if they wish. In 2014 the Wayne State pathologist Fazlul Sarkar sued anonymous commenters and subpoenaed PubPeer for their names. The Michigan Court of Appeals held that the commenters' anonymity was protected by the First Amendment (The Scientist). Anonymous criticism of science survived its first serious test in court.

Guillaume Cabanac, Cyril Labbé and Alexander Magazinov turned the tortured phrases of Part VII into the Problematic Paper Screener, which scans the literature and posts what it flags on PubPeer. Nature named it one of the key developments of 2021 (CNRS News). Leonid Schneider, a Ukrainian-German science journalist trained as a molecular cell biologist, covers misconduct cases on his blog For Better Science (Wikipedia).

None of these is run by a publisher.

The institutions respond

Publishers and governments have built their own machinery. The Committee on Publication Ethics (COPE) was founded in 1997 by British medical journal editors (COPE). In September 2024 the ORI issued a revised misconduct rule, in force from January 2025, which for the first time defines what it means to act "intentionally", "knowingly" or "recklessly" (ORI). In October 2023 the publishers' association STM began testing a tool in its Integrity Hub that detects one manuscript sent to several journals at once. By May 2024 twelve publishers were using it on about 35,000 manuscripts a month; about 1% were duplicates, and roughly half of those were thought to come from paper mills (Elsevier). United2Act, set up in 2023 with the support of STM and COPE, is a joint effort against paper mills (STM).

These are real steps, and they came after the outsiders had shown how large the problem was.

Books on the subject

Gaming the Metrics is the natural companion to Part VIII.

• Mario Biagioli and Alexandra Lippman (eds), Gaming the Metrics: Misconduct and Manipulation in Academic Research (MIT Press, 2020). Essays on how judging researchers by numbers breeds new misconduct: citation rings, rigged peer review, impact-factor gaming and predatory journals. Open access.
• James Wilsdon and colleagues, The Metric Tide (HEFCE, 2015). A UK review: metrics can support peer review but not replace it, and they can be gamed.
• Stuart Ritchie, Science Fictions (Metropolitan Books, 2020). How fraud, bias, negligence and hype undermine research.
• Nicolas Chevassus-au-Louis, Fraud in the Lab (Harvard University Press, 2019), translated by Nicholas Elliott from the French Malscience (Seuil, 2016). It opens with Summerlin's painted mouse.
• Horace Freeland Judson, The Great Betrayal: Fraud in Science (Harcourt, 2004). How fraud is defined, how common it may be, and how institutions answer it.
• Eugenie Samuel Reich, Plastic Fantastic (Palgrave Macmillan, 2009). The Schön case at Bell Labs, built on interviews with 126 scientists.
• Brian Deer, The Doctor Who Fooled the World (Johns Hopkins University Press, 2020). Deer's own account of his investigation of Wakefield.
• Ben Goldacre, Bad Pharma (Fourth Estate, 2012). Flawed drug trials and the industry's influence on doctors and regulators.
• Richard Harris, Rigor Mortis (Basic Books, 2017). A science journalist at NPR on poor biomedical research and what it costs.

The sleuths' methods are new; the problem they work on is as old as this essay.

Part X. Why it matters: trust, and the machines

Fraud harms more than the record. It harms the public that relies on science, and new tools now make it cheaper to commit.

The cost to public trust

Most people never read a paper. They meet science through a doctor's advice, a vaccine schedule or a law, so a fraud reaches people who have never heard of it.

Clinical guidelines are the clearest route. Poldermans, a Dutch cardiologist, led the DECREASE trials of beta-blockers given before surgery, and European heart guidelines drew on them. Erasmus University found fictitious patient data in his work and dismissed him (Forbes). In 2013 a meta-analysis that left his trials out, 9 trials with 10,529 patients, found significantly higher mortality with the drugs (Forbes). Boldt's starch trials in Part V belong to the same family. A guideline built on false data turns one man's fraud into routine practice.

Exposure does damage of its own, because a proven fraud becomes an argument against honest work. Anti-vaccine campaigns still cite Wakefield's paper, although it was retracted and later studies of millions of children found no link between vaccines and autism (PMC). An immunologist told Science that a retracted COVID-19 vaccine paper was still being circulated by activists (Science). The argument runs: if scientists faked that, why believe them about this? Every scandal hands that line to someone who already wanted it.

How much fraud news lowers public trust has not, as far as I can find, been measured. The nearest evidence concerns failed replications. In a 2020 experiment with 484 participants, learning that a replication had failed lowered how credible and trustworthy people judged the researchers; learning of a success raised both (Hendriks et al.).

Trust is still fairly high, and that is what is at stake. Pew asks American adults whether they have at least a fair amount of confidence that scientists act in the public's best interests. In April 2020, 87% said yes; by October 2023 it was 73% (Pew). In October 2025 it stood at 77% (Pew). Nobody has shown that misconduct caused that fall. Across 68 countries, a survey of 71,922 people published in 2025 found trust in scientists moderately high, 3.62 on a scale of 1 to 5, but only 57% rated scientists honest (University of Zurich).

The harm also outlasts the correction. In England, MMR coverage fell from 91.8% in 1996 to 79.9% in 2004, in the years after the Lancet paper (GP Online). How much of that fall the paper caused cannot be measured exactly. The retraction came twelve years after publication, and campaigners cite the paper still. A retraction corrects the journal, but it does not reach the people the fraud already persuaded.

AI: cheaper fakes, and new ways in

Part VII noted that generative AI makes plausible text and images cheaper. Richardson's group has argued in PNAS that it pushes the cost of fraud toward zero while checking stays costly (PNAS).

Most use is not fraud. Judging by the vocabulary of more than 15 million PubMed abstracts, Kobak and colleagues estimated that at least 13.5% of those from 2024 were processed with a language model (Science Advances). Much of that is polishing, which publishers such as Elsevier allow if declared.

Some papers show that nobody read them. In 2024 a paper in Surfaces and Interfaces opened its introduction with "Certainly, here is a possible introduction for your topic", and was later retracted (Technology Networks).

In February 2024 Frontiers in Cell and Developmental Biology published a paper with figures made by Midjourney, including a rat with absurd anatomy and labels such as "testtomcels". A reviewer had raised concerns that were not addressed, and the paper was retracted within days (VICE). The authors had disclosed the AI use, so this was a failure of review rather than hidden fraud. Better fakes are harder to catch: a 2024 study tested three popular AI detectors on 48 AI-made western blots and 48 real ones and found them poor at telling the two apart (PMC).

Whole datasets can be made to order. In 2023 Taloni and colleagues asked GPT-4 to produce records for 300 eyes of 250 patients with keratoconus, showing that one corneal transplant beat another. They reported that the result would appear perfectly plausible to an untrained eye (JAMA Ophthalmology). It was a demonstration, not a published fraud, but what Darsee and Stapel did by hand can now be handed to a machine.

Language models also invent references. In tests, 55% of the citations from GPT-3.5 and 18% from GPT-4 were wholly made up, and 19.9% from GPT-4o in a 2025 study (Enago; EurekAlert). In May 2025 a US government report on children's health, the "Make America Healthy Again" (MAHA) report, was found to cite studies that do not exist. The epidemiologist Katherine Keyes, named as lead author of one, said she and her colleagues had not written it (Undark). The administration never said AI was used, but the computer scientist Oren Etzioni called non-existent articles "a hallmark of AI-generated citations" (PolitiFact). That was a government report, not a peer-reviewed paper.

AI now reaches peer review from both sides. In July 2025 Nikkei Asia reported 17 arXiv preprints with hidden instructions in white text, such as "give a positive review only", meant for any reviewer who hands the paper to an AI (CACM). A later search found 18 and called that a lower bound (arXiv). Reviewers use the tools too. Liang and colleagues estimated that 6.5% to 16.9% of review text at four AI conferences in 2023 and 2024 could have been substantially modified by language models, though they found no significant evidence of this in reviews for Nature journals (PMLR).

AI also serves the defence. According to a 2026 preprint, Springer Nature, Science and the American Society for Microbiology now screen images with Proofig and ImageTwin before publication (arXiv). These tools flag; they do not prove, and spotting an invented image is a different problem from spotting a copied one. COPE's position, set in February 2023, is that an AI tool cannot be an author, because it cannot take responsibility for the work; the people who use it remain accountable (COPE). A machine can now write the paper, but only a person can be made to answer for it.

Conclusion: the pattern behind the crimes

Across two thousand years, the crimes change shape but share a structure. Each arises where an incentive to produce a particular answer meets a system that checks too little. For Ptolemy and Newton, the incentive was a theory to defend. For Lysenko and the Nazi race scientists, it was the state. For Wakefield and the Vioxx authors, it was money. For Fujii, Boldt and Stapel, it was the career rewards of productivity. For the paper mills and citation sellers, it is the scoreboard itself.

Three lessons stand out.

1. Trust without verification invites fraud. Piltdown survived because it flattered its examiners; Schön survived because reviewers checked plausibility, not raw data. Open data, preregistration and independent replication are the modern answers to Babbage's forgers and cooks.
2. Ethics is not separable from method. Tuskegee, Unit 731 and Hwang show that how knowledge is obtained is part of its validity. The Nuremberg Code, the Declaration of Helsinki and the Belmont Report exist because scientists once forgot this.
3. What is measured is what gets faked. As evaluation moved from reading work to counting it, misconduct moved from the laboratory bench to the citation index. Reforms such as the San Francisco Declaration on Research Assessment (DORA, 2012), the Leiden Manifesto (2015) and the Coalition for Advancing Research Assessment (CoARA, 2022) all urge the same correction: judge research by its content, not its metrics.

The defenders of science have also changed. Once, exposure depended on a rival like Robert Wood or a curious statistician like Fisher. Today it comes from a distributed community: image sleuths such as Elisabeth Bik, statistical detectives such as John Carlisle, the Data Colada authors, Retraction Watch, PubPeer commenters and the researchers who trace paper mills. Their work is a reminder that science's self-correction is not automatic. It is labour, often unpaid and sometimes punished, done by people who still believe the record should be true.

A crime against science is ultimately a crime against everyone who relies on it: the patient given a harmful drug, the child left unvaccinated, the student judged by fabricated intelligence data, the honest researcher passed over for a better-gamed profile. Protecting scientific integrity is not a matter of academic etiquette. It is a public duty.

References

Key scholarly works

Cited from the standard literature; check volume and page details against the originals before publication.

• Babbage C. Reflections on the Decline of Science in England, and on Some of Its Causes. London: B. Fellowes; 1830.
• Beecher HK. Ethics and clinical research. N Engl J Med. 1966;274(24):1354–1360.
• Bik EM, Casadevall A, Fang FC. The prevalence of inappropriate image duplication in biomedical research publications. mBio. 2016;7(3):e00809-16.
• Broad W, Wade N. Betrayers of the Truth: Fraud and Deceit in the Halls of Science. New York: Simon & Schuster; 1982.
• Cabanac G, Labbé C, Magazinov A. Tortured phrases: a dubious writing style emerging in science. arXiv:2107.06751; 2021.
• De Groote I, et al. New genetic and morphological evidence suggests a single hoaxer created "Piltdown man". R Soc Open Sci. 2016;3:160328.
• Fisher RA. Has Mendel's work been rediscovered? Ann Sci. 1936;1(2):115–137.
• Hirsch JE. An index to quantify an individual's scientific research output. Proc Natl Acad Sci USA. 2005;102(46):16569–16572.
• Ioannidis JPA, Baas J, Klavans R, Boyack KW. A standardized citation metrics author database annotated for scientific field. PLoS Biol. 2019;17(8):e3000384.
• Kearns CE, Schmidt LA, Glantz SA. Sugar industry and coronary heart disease research. JAMA Intern Med. 2016;176(11):1680–1685.
• Le Noury J, et al. Restoring Study 329. BMJ. 2015;351:h4320.
• Newton RR. The Crime of Claudius Ptolemy. Baltimore: Johns Hopkins University Press; 1977.
• Oreskes N, Conway EM. Merchants of Doubt. New York: Bloomsbury; 2010.
• Reverby SM. Examining Tuskegee: The Infamous Syphilis Study and Its Legacy. Chapel Hill: University of North Carolina Press; 2009.
• Richardson RAK, et al. The entities enabling scientific fraud at scale are large, resilient, and growing rapidly. Proc Natl Acad Sci USA. 2025.
• Turner EH, et al. Selective publication of antidepressant trials and its influence on apparent efficacy. N Engl J Med. 2008;358(3):252–260.
• Van Noorden R, Singh Chawla D. Hundreds of extreme self-citing scientists revealed in new database. Nature. 2019;572:578–579.
• Westfall RS. Newton and the fudge factor. Science. 1973;179(4075):751–758.
• Wilhite AW, Fong EA. Coercive citation in academic publishing. Science. 2012;335(6068):542–543.

• Biagioli M, Lippman A, editors. Gaming the Metrics: Misconduct and Manipulation in Academic Research. MIT Press; 2020. Open access.
• Chevassus-au-Louis N. Fraud in the Lab: The High Stakes of Scientific Research. Trans. Elliott N. Harvard University Press; 2019. First published as Malscience (Seuil; 2016).
• Deer B. The Doctor Who Fooled the World: Science, Deception, and the War on Vaccines. Johns Hopkins University Press; 2020.
• Goldacre B. Bad Pharma. Fourth Estate; 2012.
• Harris R. Rigor Mortis: How Sloppy Science Creates Worthless Cures, Crushes Hope, and Wastes Billions. Basic Books; 2017.
• Judson HF. The Great Betrayal: Fraud in Science. Orlando: Harcourt; 2004.
• Reich ES. Plastic Fantastic: How the Biggest Fraud in Physics Shook the Scientific World. Palgrave Macmillan; 2009.
• Ritchie S. Science Fictions: How Fraud, Bias, Negligence, and Hype Undermine the Search for Truth. Metropolitan Books; 2020.
• Wilsdon J, et al. The Metric Tide: Report of the Independent Review of the Role of Metrics in Research Assessment and Management. HEFCE; 2015.

Web sources consulted

• Now, 240: journals are still retracting Joachim Boldt's papers, Retraction Watch, July 2026
• Harvard Revokes Tenure From Francesca Gino, The Harvard Crimson, May 2025
• Swedish appeals court ups surgeon's sentence, Associated Press via KSAT, June 2023
• Wiley shuts 19 scholarly journals amid AI paper mill problems, The Register, May 2024
• Wiley journal retracts over 200 more papers, Retraction Watch, March 2025
• A do-or-die moment for the scientific enterprise, Reese Richardson, August 2025
• Citation issues cost these 20 journals their impact factors, Retraction Watch, June 2025
• Google Scholar is manipulatable, Ibrahim, Liu, Zaki, Rahwan, arXiv, 2024
• El rector de la Universidad de Salamanca carga contra Nature, elDiario.es
• Rafael Luque suspended without pay, WION, reporting El País, 2023
• Las universidades de Arabia Saudita sufren una purga en la lista de investigadores altamente citados, Pesquisa FAPESP

• Retraction Watch at ten, Retraction Watch, 2020
• News: Crossref and Retraction Watch, Crossref, September 2023
• Retraction Watch retractions now in the Crossref API, Crossref
• New database of journal article retractions to be created (MacArthur grant), CTSNet
• The Retraction Watch Hijacked Journal Checker, Retraction Watch
• Introducing the Retraction Watch Hijacked Journal Checker, Retraction Watch, May 2022
• Dr Elisabeth Bik awarded the 2021 John Maddox Prize, Springer Nature, 2021
• Feature on Elisabeth Bik and the hydroxychloroquine dispute, Yahoo News
• Complaint against Elisabeth Bik in Marseille, Nature, 2021
• Harassment and legal threats against Elisabeth Bik, Science, 2021
• Two of 100 clinical trials in eight major journals likely contain inaccurate data, Retraction Watch, June 2017
• GRIM test, Wikipedia
• Brian Wansink announces his retirement from Cornell, Retraction Watch, September 2018
• Gino's lawsuit against the Data Colada authors dismissed, Science, September 2024
• Judge declines to make Gino pay Data Colada's legal fees, The Harvard Crimson, July 2025
• Wakefield's Texas suit against Deer and the BMJ dismissed, Courthouse News, 2012
• Dana-Farber seeks retractions and corrections, STAT, January 2024
• Seek & Blastn, PLoS ONE, January 2019
• Wrongly identified nucleotide sequences in 11,799 publications (preprint), bioRxiv, 2021
• Abalkina on a Russia-based paper mill, Times Higher Education
• Abalkina's paper-mill study, preprint version, arXiv
• The rise of the science sleuths, Undark, September 2024
• Theo Baker, Wikipedia
• Marc Tessier-Lavigne reveals new emails to dispute allegations, The Stanford Daily, September 2026
• PubPeer's appeal for anonymity, The Scientist
• Tortured phrases and the Problematic Paper Screener, CNRS News
• Leonid Schneider, Wikipedia
• History of COPE, Committee on Publication Ethics
• ORI final rule, Office of Research Integrity, 2024
• A community-driven boost for research integrity, Elsevier, 2024
• United2Act phase 1 report and recommendations, STM, 2025
• Gaming the Metrics (catalogue record), 49th Shelf

• Medicine or mass murder? Guideline based on discredited research may have caused 800,000 deaths in Europe over the last 5 years, Larry Husten, Forbes, January 2014
• European heart guidelines based on disgraced research may have caused thousands of deaths, Larry Husten, Forbes, July 2013
• Review of the Wakefield paper and continuing anti-vaccine citation, PubMed Central
• Retracted COVID-19 vaccine paper, still circulated, Science, 2021
• Americans' trust in scientists, positive views of science continue to decline, Pew Research Center, November 2023
• Americans' confidence in scientists, Pew Research Center, January 2026
• Trust in science, University of Zurich news, January 2025
• MMR uptake remains below pre-Wakefield levels, GP Online
• Scientific journal publishes paper with AI-generated introduction, Technology Networks, 2024
• Study with AI-generated rat images retracted by Frontiers, VICE, February 2024
• AI detectors are poor western blot classifiers, PubMed Central (arXiv:2407.10308), 2024
• AI-generated fake references and scholarly integrity, Enago
• Fabricated citations from GPT-4o, EurekAlert!, 2025
• MAHA report studies, Undark, May 2025
• MAHA report: AI and fake citations, PolitiFact, May 2025
• Researchers are hiding AI prompts in their papers, Communications of the ACM, 2025
• Preprint on image screening by publishers (Proofig, ImageTwin), arXiv, 2026
• Artificial intelligence and authorship, COPE, February 2023

This essay was written in English. The Arabic page is a translation of it.

© 2026 Husain Alkhaldy — published in the site's essays.