Integrating ChatGPT into the ELEVATE-XR Adaptive Instructional Framework

With the fast and radical transformation of how both students and instructors can access information brought on by the development of generative AI, it is critical to develop best practices around how generative AI is used to support instruction and learning rather than being a detriment. In this paper, we examine how generative AI could be integrated into the Exercisable Learning-theory and EVidence-based Andragogy for Training Effectiveness (ELEVATE) framework developed by Stanney, Skinner, and Hughes [1] to inform adaptive approaches to instruction. Originally developed for designing eXtended Reality (XR) training experiences, the ELEVATE framework incorporates learning theories from behaviorism, cognitivism, and constructivism into a cohesive framework based on the Dreyfus and Dreyfus [2] skill acquisition model and Bloom's Revised Taxonomy [3]. The ELEVATE framework offers guidance for developing appropriate expectations and forms of instruction for students at 5 proficiency levels: novice, advanced beginner, competent, proficient, expert. The ELEVATE framework identifies language for appropriate learning objectives and types of learning activities that would be appropriate for students of different proficiency levels.