S. Mulzer, J. Williams
Generative artificial intelligence (AI) tools have changed the practical conditions under which many higher education assignments are completed. Instructors now need assignment-level guidance that is more precise than a course-wide ban or a course-wide permission statement. This paper proposes a Three-Tier Decision Framework for aligning permissible AI use with learning objectives. The framework classifies assignments by two features: (1) the cognitive complexity required by the task, using established taxonomies of educational objectives, and (2) the extent to which the task develops discipline-specific professional practice. These dimensions lead to three recommended levels of AI integration: Restricted (Tier 1), where AI use would bypass foundational learning; Guided (Tier 2), where AI can support the work under explicit instructor-defined constraints; and Open (Tier 3), where effective human-AI collaboration is part of the intended competency. The framework was refined through structured feedback sessions with 63 faculty members across four institutions and piloted with 214 undergraduate students in business and computing courses. In the pilot, students reported a clearer understanding of AI-use expectations than they had experienced under prior undifferentiated course policies, and instructors reported fewer ambiguous academic-integrity conversations. Because the pilot was exploratory and relied partly on self-report, the results should be interpreted as preliminary. The paper discusses implications for assignment design, academic-integrity communication, curriculum sequencing, and future empirical validation.
Pearl Academic Publishing. All rights reserved.
Content is licensed under a Creative Commons Attribution 4.0 License (CC-BY).
Privacy Policy | Terms of Service