Evaluating AI Output: How to Discern Accuracy, Reasoning, and Framing

Part of the Workshop Series from the AI Team at Rutgers University Libraries

Generative AI tools produce output that sounds confident and well-reasoned — but sounding right and being right are not the same thing. This workshop introduces Discernment, one of four core competencies in the AI Fluency Framework, as a structured way to evaluate AI-generated content across three dimensions: factual accuracy, quality of reasoning, and framing. Participants will practice applying this lens to real examples, moving beyond simple factchecking to identify subtler errors in logic and perspective.