Original Source
AI's Role in Science Papers Raises Concerns Over Quality Decline
AI: A Double-Edged Sword in Scientific Research
While Artificial Intelligence (AI) holds the potential to enhance the efficiency of scientific research, there are growing warnings that it could also contribute to an increase in low-quality research papers, thereby hindering scientific progress. Recent studies estimate that over 13% of biomedical paper abstracts submitted globally in 2024 were written with the aid of AI. The ease of academic paper writing enabled by AI has led to a surge in submissions, making content review more challenging and potentially fostering the spread of misinformation.
Soaring Submissions and Reviewer Burden
Key online open research repositories such as arXiv have tightened publication rules for some computer science papers in October 2025 following a surge in submissions since the advent of generative AI. Experts suggest this phenomenon has led to an increase in low-quality research. Approximately 370,000 papers were published on major platforms like arXiv, bioRxiv, and medRxiv in 2025, a 16% increase from the previous year. Professor Isao Echizen from Japan's National Institute of Informatics stated that major AI conferences are facing a shortage of reviewers amid the exponential growth in paper submissions. While AI can assist in the review process, the final evaluation still relies on human researchers, which is a time-consuming and labor-intensive task.
Debate on AI Usage and Identification Research
Debates continue among researchers regarding the extent to which AI should be used in writing scientific papers. A March 2025 survey by Nature of 5,000 global researchers found that over 90% approved of using AI for proofreading and other editing tasks, but this figure dropped to 65% for using AI in drafting papers themselves. Meanwhile, research is underway to identify AI-assisted scientific papers. In July 2025, a research team from the University of Tübingen in Germany published a report noting a significant increase in the frequency of certain words in biomedical research worldwide since OpenAI launched its AI chatbot, ChatGPT, using this to estimate the proportion of AI-generated biomedical papers.
*Source: Vietnam.vn (2026-03-28)*
