Influential study touting ChatGPT in education retracted over red flags

A research paper asserting that OpenAI's ChatGPT enhances student learning has been withdrawn nearly a year after its initial publication. Springer Nature, the journal's publisher, flagged "discrepancies" in the study's analysis and expressed doubts about its conclusions. This retraction comes after the paper gained significant traction online, amassing hundreds of citations. Ben Williamson, a senior lecturer at the Centre for Research in Digital Education and the Edinburgh Futures Institute in Scotland, highlighted that the authors made striking claims regarding ChatGPT's positive effects on learning outcomes. He noted that this study was perceived on social media as a groundbreaking piece of evidence supporting the benefits of generative AI in education. The now-retracted study aimed to measure the impact of ChatGPT on learning performance, perception, and higher-order thinking by conducting a meta-analysis of 51 previous studies. The findings reportedly indicated that ChatGPT had a substantial positive effect on learning performance, a moderate enhancement of learning perception, and encouraged higher-order thinking. Published in the journal Humanities & Social Sciences Communications on May 6, 2025, the paper's methodology faced criticism for relying on low-quality studies and for combining results from research that used vastly different methods and populations. Williamson stated that the paper should not have been published in the first place due to these methodological flaws. He also noted the implausibility of having numerous high-quality studies on ChatGPT's educational impact completed and published within just two and a half years following its launch in November 2022. Since its release, the paper garnered 262 citations in Springer Nature's peer-reviewed journals and attracted nearly half a million readers, achieving a remarkable attention score that placed it in the 99th percentile for journal articles.