Generative AI has fundamentally disrupted traditional approaches to writing assessment in higher education, rendering text-based AI detectors inadequate and ethically problematic tools for evaluating student authorship. This presentation introduces the Write–Present–Defend (WPD) model as a multimodal, theoretically grounded assessment framework that addresses AI-related academic integrity concerns while simultaneously advancing oracy, metacognition, and disciplinary communication. Drawing on behavior-analytic theory and authentic assessment pedagogy, WPD repositions integrity as an assessment design problem rather than a surveillance challenge. The model's five stages increase the effort cost of AI substitution, provide contingent feedback, and develop skills valued across generations of learners and professionals. Attendees will examine implementation tools including dual rubrics, defense question banks, behavioral indicators of authorship, and a student AI-disclosure framework. Implications for teacher preparation programs, faculty professional learning, and institutional AI governance are discussed.
