The Pedagogical Fit Of Generative Artificial Intelligence: Rethinking Assessment And Learning In Higher Education
Year: 2026
Type: article
Abstract: The rapid diffusion of generative artificial intelligence across higher education has transformed the ways in which students, instructors, and institutions conceptualize learning, assessment, and academic work. While large language models and related generative tools promise unprecedented efficiency, creativity, and scalability, their integration into educational practice has also provoked significant concerns regarding academic integrity, learning quality, cognitive dependency, and the erosion of critical thinking. The present study advances a comprehensive theoretical and interpretive investigation of generative artificial intelligence adoption in higher education by integrating task–technology fit, technology acceptance, and experience–technology alignment perspectives into a unified analytical framework. Grounded in contemporary empirical and conceptual scholarship, this research examines how generative AI tools align with the epistemic, procedural, and evaluative tasks that characterize modern higher education, with special emphasis on assessment, test automation, and behavior-driven development in digital learning environments. Particular attention is given to the implications of generative AI for automated testing and assessment design, drawing on the work of Tiwari (2025), who demonstrated how generative AI-driven behavior-driven development can enhance the efficiency, consistency, and scalability of test automation processes. This contribution is extended into the higher education domain, where assessment systems increasingly rely on automated and semi-automated tools that mirror industrial software testing architectures. Using a theoretically grounded qualitative synthesis approach, this study integrates evidence from international surveys, institutional reports, and conceptual models to interpret how students and faculty perceive, adopt, and utilize generative AI across disciplines. The findings reveal that generative AI adoption is not merely driven by perceived usefulness or ease of use but is fundamentally shaped by the degree to which these technologies fit the cognitive, disciplinary, and evaluative tasks faced by learners and educators. When task–technology fit is high, generative AI becomes an enabling infrastructure for deeper engagement, adaptive feedback, and more authentic assessment design. When misalignment occurs, however, the same technologies can undermine learning, amplify academic misconduct, and distort performance evaluation. The study contributes to theory by demonstrating that traditional acceptance models such as TAM and UTAUT must be reinterpreted in light of generative AI’s co-creative and autonomous capacities, which fundamentally alter the nature of academic tasks. It also contributes to practice by offering a nuanced understanding of how generative AI can be ethically and pedagogically embedded into higher education assessment systems. By bridging software engineering perspectives on automated testing with educational theories of learning and evaluation, this research provides a robust conceptual foundation for navigating the generative AI transition in higher education.
Author: Theodore J. Blackmoor
Institution: University of Oslo
Language: English
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: diamond
OpenAlex ID: W7218209091