Faculty are being asked to integrate generative AI into instruction faster than the empirical evidence on what works can keep up. Most institutional guidance to date has emphasized either prohibition or permission, with relatively little attention to the third and most important question: how should the learning environment be designed so that AI strengthens, rather than substitutes for, the development of professional judgment? This article is the practitioner companion to a conceptual framework on experiential intelligence in AI-mediated learning environments (Elliott, 2026). Where the framework paper argues that learning environment design determines whether students develop durable judgment or merely facility with AI outputs, this article translates that argument into specific instructional moves faculty can implement now. Drawing on recent empirical work showing that unguarded generative AI use harms long-term skill acquisition (Bastani et al., 2025; Fan et al., 2025), reduces critical thinking among knowledge workers (Gerlich, 2025; Lee et al., 2025), and intensifies automation bias in clinical decision support (Gaube et al., 2021; Kücking et al., 2024; Nguyen, 2024), the article presents twelve concrete design moves organized into four categories: (1) assignment redesign, (2) in-class practices, (3) assessment redesign, and (4) faculty development.Each move is grounded in the experiential learning tradition (Dewey, 1938; Kolb, 1984), the Dreyfus model of skill acquisition (Benner, 1984; Dreyfus & Dreyfus, 1986), and emerging literature on precision scaffolding and AI literacy in higher education (Kassorla et al., 2024). The article is intended for medical educators, faculty developers, and instructors in any discipline seeking to integrate AI without hollowing out the formation of learners. Discipline-specific adaptations are provided for medical education, theological and seminary education, and other professional disciplines.
Lydia Elliott (2026) studied this question.