"Combatting Academic Fraud and Biased Research with AI Tools and Journal Initiatives"
Originally Published 1 year ago — by Nature.com
Academic journals are increasingly using AI-based systems and expert sleuths to detect manipulated images in research papers, as concerns about image integrity continue to rise. The use of duplicate data across graphs, photo replication, and image splicing are among the issues being flagged, with some journals now requiring authors to submit raw images for review. While AI tools have proven effective in spotting certain image problems, they are less adept at detecting more complex manipulations, and as image manipulation becomes more sophisticated, addressing the issue will require broader changes in scientific culture and practices.