AI summary ‘trashed author’s work’ and took weeks to be corrected
Madison Williams-Hoffman, a doctoral student, faced issues when an AI-generated Q&A section inaccurately summarized her research on radioactive particles from Britain's atomic bomb tests, claiming it was based on only three measurements instead of 51. This error highlighted concerns about the accuracy of AI-generated content in academic publishing, with Elsevier acknowledging the mistake and removing the section after a trial, while emphasizing the potential of AI tools to assist researchers.
