Most biomedical publications show signs of LLM-assisted writing

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Lena Holzwarth, Rita González-Márquez, Dmitry Kobak
• Venue: arXiv: cs.CL
• Published: 2026-08-11
• Collected: 2026-08-12
• URL: http://arxiv.org/abs/2608.10715v1
• PDF: https://arxiv.org/pdf/2608.10715v1
• DOI: UNKNOWN
• Registry group: academic_apis_and_preprints
• Evidence role: canonical_academic_index
• Language: en
• Country / region: UNKNOWN

Plain-Language Brief

A study suggests that by the end of 2025, approximately 89% of biomedical papers on PubMed Central showed signs of LLM-assisted writing, with a higher usage in the Discussion section compared to the Methods section.

What the Document Says

• The text discusses the increasing use of Language Model (LLM) powered chatbots and agents in academic writing over several years.
• It mentions concerns about potential misconduct and fraud due to LLM-assisted writing.
• The authors propose a new unbiased method to estimate LLM usage in a corpus of texts based on changing word frequencies.
• They apply this method to open-access biomedical papers from PubMed Central for the year 2025 and find that 89% of these papers showed an excess of LLM-associated vocabulary.
• The authors also observe that LLMs are twice as likely to be used in the Discussion section compared to the Methods section, with an overall prevalence of over 50% even within the Methods section.

Key Concepts Mentioned

• LLM-assisted writing
• biomedical papers
• PubMed Central
• word frequencies

Claims Made by the Document

• By 2025, 89% of biomedical papers showed signs of LLM-assisted writing.
• LLMs are twice as likely to be used in the Discussion section compared to the Methods section.

Evidence Used

• The study applied a new method to open-access biomedical papers from PubMed Central and found an excess of LLM-associated vocabulary in 89% of these papers for the year 2025.

Methods / Approach

• The authors developed a new unbiased method based on changing word frequencies to estimate LLM usage in texts.

Limitations Stated by the Authors

The authors mention that no existing method can produce reliable estimates.

Limitations Noted by the Agent

Agent interpretation — not from the source.

The study does not provide information about the specific language model or chatbot used.

Open Questions

• What was the specific LLM or chatbot used in this study?
• How was the 'LLM-associated vocabulary' defined and identified?

Notes for Future Q&A

May include interpretation. Labeled where it does.

The text does not provide details about the specific language model or chatbot used, nor does it explicitly define what constitutes LLM-associated vocabulary.

Raw Abstract / Excerpt

Over the past several years, LLM-powered chatbots and agents have become widely used as a tool for academic writing. LLM-assisted writing can be valuable by removing language barriers but at the same time causes concerns about misconduct and fraud. To inform policy decisions, it is necessary to monitor the prevalence of LLM-altered texts in scholarly publications. Despite some recent progress in this direction, no existing method can produce reliable estimates. Here we suggest and validate a new unbiased approach to estimate LLM usage in a corpus of texts based on changing word frequencies. We apply our method to the full texts of open-access biomedical papers from Pubmed Central, and show that by the end of 2025, 89% of papers show excess of LLM-associated vocabulary. We also find that LLMs are twice as likely to be used when writing a paragraph in the Discussion section (68%) compared to a paragraph in the Methods section (32%), but even inside the Methods section, the overall prevalence of LLM usage is over 50%. We believe that our estimates are crucial to shape future guidelines and policies.