Study reveals over 70% of english biomedical papers use AI assistance, calls for regulation

A research team led by Lena Holzwarth analyzed over 1.19 million English biomedical papers from the PubMed Central platform, revealing that generative AI is rapidly becoming integrated into academic writing. The study found that approximately 77% of English biomedical papers published in 2025 showed clear signs of AI-assisted writing or editing, a significant increase from 19% in 2023 and 52% in 2024. The use of AI varied across different sections of the papers, with the discussion section showing the highest rate at around 78%, followed by abstracts at 68%, introductions at 63%, results at 58%, and methods at 54%. The study also noted that researchers from non-English-speaking countries such as South Korea and China used AI assistance at rates exceeding 80% in 2025, compared to English-speaking countries like the UK and the US. Researchers attributed this difference to the ability of AI tools to help non-native English speakers improve grammar, phrasing, and overall expression. However, the widespread use of AI in academic writing raises concerns, including the potential for generating false facts and non-existent references, as well as the tendency of large language models to present incorrect information with high confidence. Additionally, some authors failed to disclose their use of AI, which could undermine public trust in scientific research. The researchers emphasized the need for more comprehensive guidelines and regulatory frameworks rather than outright bans on AI, suggesting that clear requirements for fact-checking, reference verification, and AI disclosure are essential. They argued that AI assistance has become a common feature in academic writing, and establishing reasonable boundaries for its use is more important than simply prohibiting it.