Teaching and Learning with AI Conference

As generative AI (GenAI) tools move from experimentation to everyday use, higher education lacks practical methods for testing whether prompts perform reliably across consistent examples. This session presents a replicable prompt evaluation process using controlled multi-case testing, human rubric-based review, and revision cycles. Participants will examine ways to evaluate a GenAI prompt across varied scenarios, identify failure indicators, and iteratively refine performance by prioritizing human expertise. Together, these strategies build a practical model for evaluating GenAI integration into instructional and institutional workflows. #AI-evaluation #human-in-the-loop #prompt-fidelity