A Pedagogical Framework and Its First Classroom Implementation in Response to Automation Bias, Cognitive Debt, and the Verification Paradox

Generative artificial intelligence (GenAI) has rapidly become an embedded cognitive infrastructure in higher education, yet the pedagogical consequences of routine, uncritical reliance remain inadequately addressed in instructional practice. This paper presents the design rationale, operational specification, and first classroom implementation of the ACTIVE Framework, a six-component verification-centered pedagogical model developed at Deggendorf Institute of Technology's European Campus Rottal-Inn (DIT-ECRI), Germany. The framework responds to empirically documented phenomena including automation bias, cognitive debt, and what this paper terms the verification paradox: the tendency for students' reliance on AI to peak precisely where task complexity is highest and objective accuracy lowest, while perceived correctness remains artificially elevated. Drawing on a three-wave longitudinal pilot study (N = 21 - 36 per wave) documenting a +46.0 percentage point calibration gap at the highest complexity level the ACTIVE Framework operationalizes six principles: Awareness and Assessment, Critical Verification Protocols, Transparent Integration with Human-in-the-Loop, Iterative Skill Development, Verification Confidence Calibration, and Ethical and Contextual Evaluation, operationalized as a five-step student workflow: Assess, Constrain, Inspect, Verify, Explain. This paper reports the framework's implementation through a two-module lecture design, documents observed student failure modes in technical verification contexts, and presents an aligned assessment architecture that renders verification a teachable and gradable competency. The contribution is a context-specific, replicable instructional model developed through practitioner implementation, offered as a design narrative for educators facing similar challenges in AI-saturated learning environments.