A new study suggests that up to 30% of papers published in neuroscience journals may be fake. This worrying statistic comes as a result of the immense pressure researchers feel to publish their work, which has led to fraudulent publications being circulated. To combat this phenomenon, a team of researchers has designed a computer program to detect these attempts, with shocking results – the tool was able to pick out nearly 28% of papers as likely being made-up or plagiarized.
The fraudsters behind these papers are known as “paper mills”. They are essentially academic service providers that use artificial intelligence to scale and sell man-made works to researchers. Prices for these publications vary substantially, sometimes between $1,000 to $25,000.
The team behind the study, lead by Bernhard Sabel from Otto-von-Guericke University of Magdeburg in Germany, believe that the fake papers problem is bigger than initially thought, with paper mills continuing to grow in numbers. In response, Sabel and his team are calling for stricter procedures for reviewing publications and tougher regulations on paper mills.
Those most affected by the fakery are publishing companies and researchers. For one, publishing companies are having their reputation tarnished because of the rigged papers slipping through, which in turn could lead to decreased trustworthiness of legitimate works. Moreover, any online fraud leaves an imprint in the research community, making it more difficult for genuine researchers to progress their fields of study.
The development of AI services, like ChatGPT, can only exacerbate the situation. Indeed, it might prove difficult to keep up with the advances of fraudsters in the coming years. Nevertheless, Sabel and his team further argue that a more rigourous and collective effort from researchers, editors, and publishers could help, possibly leading to the prevention of these fraudulent papers.
