Using AI in Research and Teaching

Based on my using AI in historical research, I totally agree with Professor David Bell that AI “can produce a B-level paper. It can’t produce an A-level one” and Rahul Kalavagunta that AI “is not useful without precise prompts.” In response to Peter Severson’s analogy of gym exercise, I believe that AI can help everyone become a B-level athlete, but only you yourself can prompt AI to help you become an A-level one. To make “precise prompts” requires a lot of expertise skills and intellectual exercises (e.g. always asking “What’s the right way to solve this problem,” in Tom Griffiths’ words) that shall make you smarter. To use AI generally is one thing, to use it intellectually and expertly is another, for which we must encourage and train our students. My method is to let AI answer my question first, and then ask the students to criticize and raise their own questions for AI to answer until the students satisfy. I grade the students’ criticisms, questions, and the final answers (which actually are their collaboration with AI).