Before, During, and After AI: Redesigning Assessment to Measure Metacognitive Performance and Learner Judgment with UnBlooms™

Research shows AI can produce a strong assignment without necessarily building the learner’s capacity. When educators assess only the final product, they cannot tell whether AI extended the learner’s reasoning, substituted for it, or left anything transferable behind.

This interactive session introduces UnBlooms™, a recursive, non-hierarchical framework for redesigning assessment around evidence of learner thinking before, during, and after AI use. Through the practical Question–Generate–Critique–Refine cycle, participants will explore how to establish an independent human baseline, interrogate AI-generated evidence and assumptions, and require learners to explain and defend their decisions. The performance-based Metacognitive Awareness Scale and AI-free transfer tasks offer ways to assess observable changes in learner judgment and determine what persists beyond AI-assisted performance.

Participants will see how the performance-based Metacognitive Awareness Scale, decision trails, first-pass acceptance, misconception audits, and AI-free transfer tasks can make changes in learner judgment observable over time. Attendees will leave with practical ways to establish an unaided baseline, require learners to interrogate AI-generated evidence and assumptions, and assess what they can explain, defend, create, or resist without assistance.