Springer Nature retracts influential study claiming ChatGPT boosts student learning
Despite accumulating hundreds of citations and half a million readers, the paper is now deemed invalid by the journal publisher
Springer Nature has formally retracted a highly cited study published in May 2025 which claimed that OpenAI's ChatGPT positively impacts student learning outcomes. The retraction notice, issued on April 22, 2026, cites discrepancies in the meta-analysis and a lack of confidence in the validity of the analysis and resulting conclusions.
The retracted paper attempted to quantify the effect of ChatGPT on students' learning performance, learning perception, and higher-order thinking by analysing results from 51 previous research studies. The analysis supposedly showed that ChatGPT has a large positive impact on improving learning performance along with a moderately positive impact on enhancing learning perception and fostering higher-order thinking.
Despite the withdrawal, the paper had already garnered 504 citations and attracted nearly half a million readers prior to the notice being widely shared. The study was first published in the journal Humanities & Social Sciences Communications and claimed to measure the effect size between experimental groups that used the AI chatbot and control groups that did not.
Experts, including Ben Williamson from the University of Edinburgh, had previously questioned the study's methodology and timing. Williamson noted that the paper synthesised poor-quality research and mixed findings from studies with incompatible methods, populations, and samples. He also highlighted that it was infeasible for dozens of high-quality comparative studies to have been conducted, reviewed, and published in the short window since ChatGPT's 2022 launch.
The authors of the retracted paper did not respond to correspondence from the journal editor regarding the concerns raised. Williamson warned that the headline finding that ChatGPT improves learning may persist in the public consciousness despite the retraction, as social media circulation stripped away the nuances of the study's flaws.
The retraction comes as educators continue to grapple with adapting classrooms to prevent AI-enabled cheating while tech companies promote chatbot tools as study aids. The decision underscores the challenges of interpreting rapidly evolving research in the field of artificial intelligence and education.
