他因拷问问题论文而声名鹊起
Misconduct sleuth in China swiftly gains acclaim for calling out questionable papers
https://www.science.org/content/article ... id=5981465
中国在应对猖獗的科研不端行为方面，获得了一位意想不到的盟友：一名独立的视频博主，他向数百万粉丝揭露问题论文。
自今年4月以来，北京航空航天大学前生物医学博士生耿红伟，一直在揭露中国一些顶尖高校知名科学家的论文中明显存在的图片重复使用和可疑数据。他在施普林格·自然集团出版的九篇论文中发现的问题，已引发多次官方调查，并导致数名高层学者被解职。
观察人士表示，对耿红伟的发现以及当地媒体对曝光丑闻的广泛报道所做出的迅速回应，表明官方表现出不同寻常的容忍，甚至可能是认可。
“这一事件具有里程碑意义，”黄冈师范学院研究科研诚信的学者徐绍雄表示，“这很可能是草根举报人利用社交媒体，首次在中国官方媒体上获得广泛正面报道。”
中国的科研不端问题已不是秘密。北京大学神经科学家饶毅表示，过去20年，中国科学论文的数量和质量都有所提高。但他补充说，不仅包含造假行为的论文数量在增加，其比例也在上升，“创下了没有哪个中国人会感到自豪的历史记录”。他说，尽管国家和个人都做出了努力，但即使是众所周知的学术不端案件也未被追究。
耿红伟误打误撞地成为了一名学术不端行为的调查者。读研期间，他创办了视频博客频道“耿同学讲故事”。这些视频面向研究生同学，内容涵盖科研进展，特别是生物医学领域的进展、实验室生活轶事以及对学术不端行为的评论。2025年5月，由于对研究前景不满意，他放弃了博士学位，开始全职从事视频博客创作。
耿红伟偶尔制作关于学术不端的视频，这为他带来了更多线索。他意识到自己在B站和抖音上的300万粉丝让他拥有了超越普通举报人的影响力。在接受《科学》杂志采访时，他表示自己决定瞄准知名科学家的不端行为，因为他们获得了大量资助，应该对此负责。他关注网上帖子中标记的可疑图片和数据，并开始仔细查阅公开的原始数据，寻找异常。例如，在一项研究人员记录了小鼠体重的研究中，耿红伟发现数字“5”出现的频率异常高，而非预期的随机数字分布。“这显然不是实验室通常产生的那种数据，”耿红伟说。
在过去两个月里，耿红伟提供了涉及中国知名科学家的九篇生物医学论文中存在图片和/或数据操纵的证据。这些论文都发表在施普林格·自然的期刊上。他表示，这些期刊受到中国顶尖科学家的青睐，并且提供更多原始数据，便于进行更严格的审查。（《自然》研究与综述期刊副总裁埃里卡·帕斯特拉纳在给《科学》杂志的一封电子邮件中写道：“我们正在严格调查提出的这些问题，目前这些仍处于评估阶段的指控。”）
大多数涉事机构已启动调查。同济大学、南开大学和中山大学已经解雇或处分了被认为对不端行为负责的基层研究人员，并因监督不力而对包括院长级学者在内的高级科学家进行了降职或解职。
日本广岛大学的高等教育学者黄福涛表示，快速的调查和纪律处分表明，当局认识到科研诚信“已与中国的国际科学声誉及其成为全球科学领导者的雄心紧密相连”。这些回应与过去的做法形成鲜明对比。例如，2021年，中国科技部调查了时任南开大学校长曹雪涛参与合著的60多篇论文中存在的图像操纵指控。尽管发现“多篇论文存在图像使用不当，反映实验室管理不严”，但科技部得出的结论是没有欺诈证据，曹雪涛仅受到轻微制裁。
耿红伟表示，他得到的公众反馈，甚至与政府官员的有限接触，基本都是支持的。他认为，他从年轻研究人员那里收到的几十条线索表明，他们对当前的研究实践存在“非常明显的不满”。他说，大量炮制有问题的论文“对这个国家的科研发展非常有害”。他告诉《科学》杂志，他目前正在调查大约20篇可疑的出版物。
他说，减少数据滥用的一个方法是让研究团队的其他成员独立重复关键实验。“如果我知道有人会重复（这个实验），我就不敢伪造数据。”
徐绍雄说，中国依赖像耿红伟这样的博主来曝光不端行为，暴露了机构监督机制的失灵。在最近的一篇论文中，他和同事发现，在579所中国大学中，只有52.8%的大学维护着学术诚信网页，仅有16.8%的大学按照教育部的要求发布了年度科研诚信报告。
黄福涛说，激励机制加剧了学术不端问题：对教师和学生的评价应减少对论文数量和期刊声望的依赖，而更多地关注研究的原创性、可重复性和长期科学价值。大多数人认为，被判定犯有不端行为的人也应面临更严重的后果。正如饶毅所说：“应该有真正有效的机制和真正有责任感的官员来适当、公平地调查研究不端行为，并采取适当的纪律处分。”
本文由Bian Huihui参与报道。
doi: 10.1126/science.zdqjrzj
关于作者
丹尼斯·诺米尔报道亚洲（尤其是中国和日本）的科研与科学政策动态。他在东京工作。
Independent vlogger Geng Hongwei galvanizes often-sluggish institutions to investigate and punish prominent scientists
China has gained an unexpected ally in its struggle with rampant scientific misconduct: an independent video blogger who shines a light on problematic papers for millions of followers.
Since April, Geng Hongwei, a former biomedical Ph.D. student at Beihang University, has been uncovering apparent use of duplicated images and questionable data in papers authored by prominent scientists at some of China’s premier universities. Problems he has identified in nine papers published in Springer Nature journals have triggered numerous official investigations and the dismissal of several high-ranking academics.
Observers say the swift response to Geng’s findings as well as the extensive local media coverage of the unfolding scandals indicate an unusual level of official tolerance, if not approval.
“This case is a landmark event,” says Shaoxiong Brian Xu, a research integrity scholar at Huanggang Normal University. “It is likely the first time a grassroots whistleblower utilizing social media has achieved widespread, positive coverage in Chinese state media.”
China’s scientific misconduct problem is no secret. The number of scientific papers from the country has increased over the past 2 decades, as has the quality of those publications, says Yi Rao, a neuroscientist at Peking University. At the same time, not only the number, but also the proportion of papers containing fabrications has increased, he adds, “[setting] historical records that no one from China is proud of.” Despite efforts at the national level and by individuals, even well-known cases of misconduct have gone unpunished, he says.
Geng more or less stumbled into his role as a misconduct sleuth. As a graduate student, he started the video blogging channel “Classmate Geng Tells Stories.” Aimed at fellow graduate students, the videos covered research developments, particularly in biomedicine; anecdotes about life in the lab; and comments on academic misconduct. In May 2025, unhappy with his prospects for a research career, he quit his Ph.D. studies and began vlogging full time.
Geng’s occasional videos on misconduct attracted further tips, and he realized his 3 million followers on the video platforms Bilibili and Douyin gave him visibility beyond that of the typical whistleblower. In an interview with Science, he said he decided to target misconduct by high-profile scientists because they receive generous funding and should be held accountable. He watched for online posts flagging questionable images and data and started scouring publicly accessible raw data, looking for anomalies. For example, in a study where researchers had recorded the weight of mice, Geng spotted an unusual preponderance of fives instead of the random frequencies of digits that would be expected. “This clearly wasn’t the kind of data normally generated in a laboratory,” Geng says.
Over the past 2 months, Geng has presented evidence of image and/or data manipulation in nine biomedical papers involving prominent Chinese scientists. All were published in Springer Nature journals, which Geng says are favored by top Chinese scientists, and which make more raw data available, allowing for greater scrutiny. (In an email to Science, Erika Pastrana, Vice President of Nature Research and Reviews Journals, wrote, “We are rigorously investigating the concerns that have been raised, which at this stage remain allegations under assessment.”)
Most of the institutions involved launched investigations. Tongji University, Nankai University, and Sun Yat-sen University have already dismissed or disciplined the bench-level researchers judged responsible for the misconduct and have demoted or dismissed senior scientists, including dean-level scholars, for lax oversight.
The quick investigations and disciplinary actions suggest authorities recognize that research integrity “has become closely linked to China’s international scientific reputation and its ambition to become a global scientific leader,” says Futao Huang, a higher education scholar at Hiroshima University in Japan. The responses contrast with past practice. In 2021, for instance, China’s Ministry of Science and Technology probed allegations of image manipulation in more than 60 papers co-authored by Cao Xuetao, then-president of Nankai. The ministry concluded there was no evidence of fraud despite finding “misused images in many papers, reflecting a lack of rigorous laboratory management,” and Cao faced only minor sanctions.
Geng says feedback he has received from the general public and even from limited contacts with government officials is generally supportive. He believes the dozens of tips he gets from younger researchers indicates “very clear dissatisfaction” with current research practices. Churning out flawed papers is “very detrimental to the development of scientific research in this country,” he says. He told Science he is now looking into 20 or so suspicious publications.
One way to reduce data misuse, he says, would be to have other members of a research team independently replicate key experiments. “If I know that someone will replicate [the experiment], I wouldn’t dare to fabricate data,” he says.
China’s reliance on a blogger like Geng to expose misconduct reveals a breakdown in institutional oversight, Xu says. In a recent paper, he and his colleagues found only 52.8% of 579 Chinese universities maintained academic integrity web pages, and just 16.8% released annual research integrity reports, as required by the Ministry of Education.
Incentive structures contribute to the misconduct problem, Huang says: Evaluations of faculty and students should rely less on publication counts and journal prestige and more on research originality, reproducibility, and long-term scientific value. Most agree that those found guilty of misconduct should also face greater repercussions. As Rao says, “There should be mechanisms with real teeth and officials with a true sense of responsibility to properly and fairly investigate research misconduct and take proper disciplinary actions.”
With reporting by Bian Huihui.
doi: 10.1126/science.zdqjrzj
About the author
Dennis Normile writes about research and science policy developments in Asia, particularly China and Japan. He is based in Tokyo.
