An influential scientific paper claiming a positive impact of ChatGPT on student performance has been retracted less than a year after its publication, reports Todayinfo.
The prestigious publisher Springer Nature made this decision due to «inconsistencies» in the analysis and a lack of confidence in the conclusions. However, by then, the article had already been cited hundreds of times and widely circulated on social media.
This research was based on a meta-analysis of 51 other works. But experts criticized it from the outset. According to them, the authors combined low-quality or methodologically unsuitable studies, leading to erroneous conclusions.
It was deemed impossible to conduct and publish so many high-quality studies in such a short period after ChatGPT's release.
The most dangerous aspect is not the study itself, but its consequences. On social media, only the headline «ChatGPT is beneficial for education» spread, while its methodological flaws were ignored. As a result, this erroneous assertion became the basis for many other scientific works and contributed to the spread of misinformation.
