Great start to our doctoral seminar for first-year PhD students. We discussed Richard Smith’s “Peer Review: A Flawed Process at the Heart of Science and Journals.”
I was struck by how quickly the conversation turned to the operational dimensions of the problem, while rarely engaging with its human aspects–or even with the author’s underlying cynicism. Given that, several enthusiastic proposals quickly emerged about using AI to conduct peer review more objectively and efficiently.
I was tempted to propose an experiment for the coming weeks: to split the class into two groups—one in which students would discuss the assigned texts only on the basis of AI-generated comments in sessions led by a chatbot, and another in which students would engage only through human interaction in sessions led by myself. Then, to observe what happens. And, at the end of the semester, ask the students which approach they prefer.