Workflow
Nature头条:学术期刊正在被AI撰写的“洗稿论文”攻陷
生物世界·2025-09-24 08:30

Core Viewpoint - The rise of generative AI, such as ChatGPT, has led to an increase in low-quality, redundant academic papers, which are infiltrating reputable journals and bypassing plagiarism detection systems [3][6][7]. Group 1: Research Findings - A study from Surrey University revealed that between 2021 and July 2025, 411 redundant papers were published across 112 academic journals, utilizing AI tools like ChatGPT to create "copycat" papers [6][9]. - The research specifically focused on studies using the NHANES dataset, highlighting that some papers were published multiple times, with one study on oxidative balance and chronic kidney disease appearing six times within a year [6][7]. - The study indicated that 37% of the redundant papers were published in journals under Springer Nature, with 51 papers in Scientific Reports, which is now investigating these cases [9]. Group 2: Implications for Academic Publishing - The ability of AI to generate new papers from existing data poses significant challenges for academic journals, as these AI-generated papers can evade traditional plagiarism checks [8][9]. - Since the public release of ChatGPT in 2022, there has been a notable increase in redundant research papers based on NHANES data, prompting some journals to tighten their submission policies regarding such studies [8][9]. - The findings suggest a potential overwhelming of genuine academic literature by AI-generated content, raising concerns about the integrity of scientific publishing [7].