Practicing Class Presentations with AI

Presentations ask a student to communicate an idea to listeners who cannot reread every sentence. Cornell’s Center for Teaching Innovation recommends practice opportunities, rubrics and feedback for student presentations, including rehearsing part of a talk with peers. AI can help organize that rehearsal, but it does not see every audience reaction or know the instructor’s criteria unless provided. Start with the time limit, topic, audience, allowed aids and rubric. Build a brief outline with a point for each section and a reason the audience should care. Ask AI to identify a missing transition, suggest a question a listener might ask or check whether a slide has too much unexplained detail. Verify every factual claim and chart label against the source. A generated analogy can be memorable yet misleading. Keep slides readable and provide any captions or descriptions required for access. The tool can help draft alternatives, but the presenter chooses what they can explain accurately. Rehearse aloud. Time the introduction, one difficult section and conclusion, then the whole talk. A model may evaluate a transcript, but transcript quality cannot measure eye contact, projection or how listeners respond. Record a practice run where permitted, or ask a peer for feedback. Compare comments with the actual rubric rather than accepting a generic score. Prepare for questions by having AI generate a few plausible challenges and practicing a concise, source-grounded answer; it is acceptable to say what remains unknown. After rehearsal, change one high-impact issue at a time and practice again. Avoid replacing the student’s voice with a script they cannot own under questioning. Follow the course’s AI-use rules, especially for graded content or uploaded recordings. A useful AI rehearsal partner sharpens structure and prompts practice, while the student and human audience determine whether the talk communicates.