Colloquial engagement theory with AI awareness (CET-AIA): A new creative pedagogical framework for ethical assessment in the age of artificial intelligence

• Introduces CET-AIA, a new framework for ethical AI-resistant assessment. • Embeds colloquial, context-rich questioning to reduce AI-assisted cheating. • Empirical tests show students adapt while AI performance declines sharply. • CET-AIA enhances critical thinking, engagement, and assessment reliability. • Provides scalable, creative strategies for future AI-aware pedagogy. This study introduces the Colloquial Engagement Theory with AI Awareness (CET-AIA), a novel pedagogical framework developed to address academic integrity challenges in higher education posed by generative AI systems. CET-AIA integrates informal, context-sensitive questioning into assessment design to discourage AI-assisted cheating and enhance authentic student engagement. A quasi-experimental design was implemented across three undergraduate courses, comparing student and AI (ChatGPT-3.5) performance on traditional versus colloquial multiple-choice assessments. Theoretical foundations were drawn from Constructivism, Cognitive Load Theory, Sociocultural Theory, and Authentic Assessment. Performance trends were analyzed using t-tests, ANOVA, regression models, and mixed-effects modeling. Results indicate a significant initial decline in student scores under colloquial questioning, followed by gradual improvement, confirming both the cognitive challenge and adaptation process. In contrast, AI models showed a persistent performance drop when confronted with colloquial, context-specific questions. These outcomes demonstrate CET-AIA’s effectiveness in fostering deeper learning and in shielding assessments from AI exploitation. CET-AIA offers a scalable framework for designing AI-resistant assessments that promote critical thinking and real-world comprehension. The approach aligns with current educational priorities in fostering academic honesty and student-centered learning in digitally enhanced environments. This research is among the first to offer a theoretically grounded, empirically validated framework specifically designed to neutralize AI-assisted cheating through linguistic and contextual innovation. CET-AIA bridges the gap between AI ethics, pedagogy, and assessment, presenting a future-ready model for ethical education.