AI Agents Struggle to Interpret Science, Researchers Propose Strategies

C&EN reports on a study by Valentin Rodionov revealing AI agents' limitations in distinguishing solid science from fraudulent studies, prompting calls for improved strategies.