AI agents scan scientific literature and catch decades-old errors

AI agents are proving adept at scanning scientific literature and reference databases to identify errors in papers published decades ago. The technology's ability to flag long-standing mistakes could have significant implications for the integrity of scientific research.