Research yet to be peer reviewed suggests role for tools like ChatGPT in REF processes
Generative artificial intelligence has been found to perform better than citation data for assessing the quality of academic papers, suggesting a potential use for AI in the Research Excellence Framework, according to a research project in its early stages.
Mike Thelwall, a professor of data science at the University of Sheffield, is leading an Economic and Social Research Council–funded project exploring whether large language models such as ChatGPT could help assess the research quality of academic journal articles—and his initial findings suggest they can.
