How AI Simulations Force Productive Failure

This conversation explores a research proposal that outlines a multi-course investigation led by Jonathan H. Westover and colleagues at Utah Valley University to test AI-enabled immersive simulations across diverse management disciplines. Building on a successful summer 2026 pilot in a single leadership course, the study expands across thirteen course sections in fall 2026 and spring 2027 to examine the knowing-doing gap in management education. The authors propose a theoretical process model wherein consequential simulation tasks create productive struggle, revealing gaps between students' self-perceptions and actual performance through detailed career profile metrics. Furthermore, the research investigates how diagnostic feedback and structured reflection facilitate mental-model revision, provided that students maintain trust in the algorithmic assessment. Ultimately, this expanded initiative aims to evaluate how task characteristics—such as ambiguity, rule boundedness, and relational complexity—influence whether students develop adaptive judgment and successfully transfer their learning to novel professional scenarios.