Retractions

A test of whether an AI pre-submission review would have pointed at the documented problem in 23 open-access papers retracted for errors visible in the manuscript. It did in 14, partly in 4, and missed 5.