Despite decades of empirical research in multimedia learning, many higher ed videos often suffer from
"Death by PowerPoint"—cluttered, cognitively taxing, hindering learning and learner engagement. This session introduces the
Multimodal Video Evaluator (MVE), a custom AI bot designed to close the research-practice divide by translating scientific research into an automated, systematic pedagogical auditor. The MVE codifies a 15-year synthesis of the
Cognitive Theory of Multimedia Learning (CTML) and
video engagement studies into
127 actionable design guidelines.
By replacing pedagogical guesswork with a systematic approach to rapid design-revision cycles, the MVE empowers educators to systematically transform passive, distracting slides into high-impact visual tools that significantly improve student comprehension, memory retention, and learner engagement.
To foster institutional adaptation, I will reveal the MVE’s
full design architecture, including its specialized knowledgebase and system prompt. I will discuss
calibration features used to ensure accurate and consistent auditing and prevent hallucinations of generic AI. By sharing these internal blueprints, I provide attendees with the technical and pedagogical insights needed to replicate and refine this work at their own institutions, moving AI from an efficiency tool to a framework for
quality assurance and
competency building for faculty, instructional designers, and students.
My Substack Publication "eLearning Expert": https://elearningexpert.substack.com/ My LinkedIn: https://www.linkedin.com/in/k0ichisat0/
