Taiwan's Academic Ghostwriting Market Exposes a Trade in Papers, Authorship and Publication

A NT$470,000 offer behind the pressure to publish

A retired teacher from a private university in Taiwan, identified only as Teacher A, described encountering an international paper mill charging about NT$470,000 and arrangements in which journal editors allegedly helped secure publication. Her account, dated October 12, 2026, also described a separate NT$30,000 ghostwriting purchase that reportedly turned out to contain text generated indiscriminately by artificial intelligence.

The two prices capture different parts of an academic underground market: inexpensive writing sold online and much more costly packages involving the machinery of scientific publication. They are individual accounts, not established market averages. The identities of the sellers, the journals involved and any resulting investigations were not disclosed.

Teacher A, who became an informal academic adviser online after retirement, portrayed some customers as teachers worn down by applications for National Science and Technology Council funding and university evaluations. In her account, unequal power relationships and professional insecurity helped drive demand. Buyers were seeking reassurance as well as written material or research graphics.

Taiwan's official ethical guidance draws a firm distinction between assistance and substitution. Having a colleague or professional editor improve a manuscript can be legitimate. Having someone who did not participate in the research write the entire paper is not. That distinction matters because the underground trade includes several different transactions, from thesis writing to purchased authorship and alleged interference with journal review.

The account does not establish how widespread these practices are in Taiwan. It does, however, describe conduct that connects local academic pressures with an international market whose products can enter the scientific record and compete for public research funding.

Where editing becomes academic misconduct

The National Science and Technology Council's academic ethics explanation, on a page updated July 21, 2015, explicitly addresses the boundary between ghostwriting and acceptable editorial help. It says complete writing by an outsider who did not participate in the research is improper, while revision and polishing by peers or professional editors are common legitimate practices.

The document also acknowledges that cases between those endpoints can be difficult to judge. Its stated assessment standard focuses on whether conduct is misleading and serious. It cautions against treating an ordinary mistake as an ethical violation capable of destroying a researcher's reputation.

This is a more precise distinction than simply asking whether money changed hands. Paying for language editing does not by itself establish misconduct. Conversely, a manuscript can be problematic even if it passes a plagiarism check, because the named author may not have performed the research or written the work being presented as their own.

The guidance covers research ideas, execution and presentation of results. It addresses data collection, preservation of research records, acknowledgement of contributions, authorship, confidentiality in peer review and disclosure of conflicts of interest. It also identifies fabrication, falsification, plagiarism and improper publication practices as unacceptable.

These categories help explain why a paper mill transaction can involve more than ghostwriting. Invented results concern fabrication; altered images or data can involve falsification; purchased author positions concern the truthful attribution of contributions. A single paper may present several distinct ethical problems.

What the enforcement record actually shows

The council's historical figures record 98 academic ethics cases received between 1999 and 2012. Seven involved fabrication. Most of the remaining cases concerned plagiarism or inadequate citation.

Seven out of 98 is approximately 7.1%. That calculation describes the composition of cases received during that period, not the prevalence of fabrication among Taiwanese researchers. The figures also do not isolate ghostwriting or provide a count of paper mill transactions.

According to the same official explanation, sanctions ranged from written warnings to suspension of eligibility for one to ten years, with one case resulting in lifetime suspension. These are historical enforcement figures, not a statement that every form of ghostwriting automatically receives a particular penalty.

A separate account of proposed amendments under then Minister of Science and Technology Chen Liang-gee described plans to classify ghostwriting expressly as research misconduct. Chen said researchers with ghostwritten papers could be required to return government funding if the proposed rules were approved.

That account also reported 11 researchers punished from January through June: eight were barred from funding applications for one to five years, and three received warnings. One researcher was required to return NT$600,000. The repayment was an additional consequence, not a twelfth case. The year of that reporting period is not specified, so the figures cannot responsibly be presented as a current annual total.

A changing rulebook, with limits on what is confirmed

The council's academic ethics portal lists guidance for researchers, an academic ethics statement revised July 28, 2022, and specific guidance for research projects involving student theses and dissertations. It also lists rules for handling and investigating misconduct marked as revised February 6, 2026.

That listing establishes the existence and revision dates of the documents. It does not, by itself, establish the exact wording of the 2026 rules, whether every earlier proposal was adopted, or which provisions would apply to Teacher A's allegations.

The earlier amendment account described investigations generally reaching back ten years, with exceptions involving continuing funding or threats to public interests or safety. It also described an institutional investigation followed by ministry review through a task force of at least three relevant specialists. Serious cases would receive further examination by an ethics review committee. Those details should be understood in their reported historical context, rather than assumed to reproduce the latest rules.

The documented sequence separates older enforcement evidence from later policy developments:

• 1999 to 2012: The council received 98 academic ethics cases, including seven involving fabrication.
• July 21, 2015: The official explanation distinguishing complete ghostwriting from legitimate editing was updated.
• July 28, 2022: Several ethics documents listed on the council portal were revised.
• January 2, 2024: A study of 102 ghostwriting websites was published in Higher Education.
• February 6, 2026: The portal marks the misconduct handling rules as revised.
• October 12, 2026: The account describing Teacher A's encounters was dated.

How ghostwriting sellers make misconduct sound acceptable

A study by Kenneth Han Chen and John Chung-En Liu, published in Higher Education on January 2, 2024, examined the marketing language of 102 academic ghostwriting websites in English and Chinese. Their ghostwriting marketing study investigated how businesses justify services that carry an ethical stigma.

The authors identified five techniques: denial of responsibility, denial of injury, denial of the victim, condemnation of the condemners and appeal to higher loyalties. In plain language, these are ways of shifting blame, presenting the transaction as harmless, disputing that anyone is wronged, attacking critics or claiming that some other obligation justifies the conduct.

The study found a difference in emphasis across the two language markets. Chinese websites stressed the authority of their ghostwriters, while English websites gave greater weight to empathy with users. The finding concerns marketing strategies, not proof that every customer in either market responds in the same way.

That distinction helps interpret the Taiwanese accounts. Legislator Hung Tzu-yung previously described thesis ghostwriting costing about NT$120,000 and said students could request a professor or a person with a master's degree as the writer. Selling the writer's credentials turns academic status itself into part of the product.

Teacher A's description of anxious academics seeking reassurance fits another aspect of the market: services can be presented as relief from institutional pressure. Her account and the website study do not establish that pressure causes every purchase. Together, however, they show how sellers can appeal both to confidence in a supposed expert and to a customer's sense of vulnerability.

Three prices, three different transactions

The Taiwan accounts describe NT$30,000 for an online writing purchase, about NT$120,000 for thesis ghostwriting cited in legislative discussion, and about NT$470,000 for the international paper mill offer encountered by Teacher A. The highest figure is roughly 15.7 times the lowest, but the services and reporting periods differ. This is not a valid measure of price inflation.

International accounts describe authorship services priced at approximately €180 to €5,000, with reported dollar equivalents of US$197 to US$5,472. These services can place a customer's name on an article without that customer carrying out the research. Some businesses sell positions before acceptance; others market positions when a paper is ready for publication.

The distinction between buying text and buying authorship matters. A thesis ghostwriter may produce work submitted for a degree. A paper mill may sell a position on a journal article, together with a promise of publication. The second transaction can misrepresent both the contribution of the named author and the reliability of the publishing process.

These prices cannot be combined into a dependable estimate of Taiwan's market size. There is no disclosed count of customers or transactions, and the accounts do not establish what each package included. The figures demonstrate a range of reported commercial arrangements, rather than a single standardized service.

When the publication gatekeepers become part of the scheme

International reporting describes paper mills using copied text, false or stolen data, manipulated images, artificial intelligence rewriting and fabricated peer reviews. Some also offer guaranteed publication. A promise of acceptance before genuine editorial assessment is a warning sign because a legitimate decision depends on evaluating the manuscript, usually with independent reviewers.

Peer review is the process in which specialists assess a paper's methods, evidence and reasoning before publication. Fake reviews imitate that safeguard without supplying its independent scrutiny. When an editor cooperates with a seller, the problem extends beyond the manuscript to the journal's decision making.

A reported investigation in 2024 identified more than 30 journal editors involved in paper mill activities and described offers to editors of as much as US$20,000. Those international findings provide context for Teacher A's allegation of coordination with journal editors. They do not independently prove the particular arrangement she described.

Other reported warning signs include authors with unrelated specialties, undisclosed target journals and customers who do not know their supposed coauthors. Unusual collaboration alone is not proof of fraud. Its significance depends on whether the listed authors actually contributed and whether the research and publication records support the claimed work.

Retractions show the damage, not the full market

Reported figures for retractions linked to paper mills rose from ten in 2019 to 2,099 in 2023, about 210 times the earlier number. Hindawi and Wiley reportedly retracted approximately 1,200 such articles in 2023. That publisher figure belongs within the wider problem and should not simply be added to the 2,099 total.

As of May 2024, Retraction Watch had recorded 7,275 retractions associated with paper mills among approximately 44,000 retractions in its database. That is about 16.5% of recorded retractions, not 16.5% of all scientific papers. Estimates that as many as 400,000 paper mill articles may have entered the literature over two decades are much less certain than documented withdrawal counts.

Retraction means a journal has withdrawn an article from the reliable published record. Rising numbers can reflect greater detection and large batches of withdrawals, as well as fraudulent output. They cannot establish on their own that production increased at the same rate.

Financial estimates also require care. The UK Research Integrity Office has been cited as estimating approximately US$10 million in global industry gains. A separate calculation suggested one Russian mill could have earned US$6.5 million if it sold all authorship positions on papers it produced from 2019 to 2021. The latter is a conditional estimate, not confirmed revenue, and the two figures are not directly comparable.

AI checks cannot settle the authorship question

The reported NT$30,000 purchase that turned out to contain indiscriminately generated AI text illustrates a practical weakness in buying academic work: customers may receive material that cannot support the claims made in their names. The account says the problem emerged after a plagiarism check, but does not identify the checking tool or explain how AI use was established.

A plagiarism check compares text with other material to identify overlap. It does not establish that an experiment happened, that data are authentic or that a listed author made the claimed contribution. Text can have little detectable overlap and still be fabricated or written entirely by an outsider.

The council's guidance on preserving research records, acknowledging contributions and maintaining reliable data therefore addresses questions that a text similarity score cannot answer. The central issue is whether the work and its authorship are honestly represented, not merely whether the wording appears original.

No named institution, journal or seller has been identified in Teacher A's account, and no hearing, investigation deadline or formal finding is disclosed. The evidence supports scrutiny of the reported practices, while leaving their scale in Taiwan and the outcome of the specific allegations unresolved.

Key Points

• A retired Taiwanese teacher described a NT$470,000 international paper mill offer and alleged coordination with journal editors.
• A separate NT$30,000 writing purchase reportedly contained indiscriminately generated AI text.
• Official Taiwanese guidance distinguishes legitimate manuscript editing from complete ghostwriting by someone outside the research.
• The council received 98 ethics cases from 1999 to 2012; seven involved fabrication.
• A study of 102 websites found Chinese ghostwriting marketing stressed writer authority, while English marketing stressed empathy.
• Reported paper mill retractions rose from ten in 2019 to 2,099 in 2023, but withdrawal counts do not measure the full market.
• The council portal lists misconduct handling rules revised in February 2026; the specific allegations have no disclosed formal outcome.