Rethinking Student Assessment in the Era of AI-Assisted Learning

Abstract

The rapid integration of artificial intelligence (AI) technologies in higher education is transforming teaching, learning, and assessment practices. Generative AI systems, such as ChatGPT and Google Gemini, enable students to generate ideas, summarize academic materials, and refine written work, challenging the validity of conventional assessments designed to measure independent intellectual effort. Despite this potential, limited research has explored how assessment can be reconceptualized to accommodate AI-assisted learning while maintaining academic rigor. This study addresses the question, how student assessment can be aligned with AI-mediated learning in higher education. Guided by Extended Mind Theory, which conceptualizes cognition as distributed across human and technological agents, the study examined assessment in Tanzanian higher learning institutions. Using a qualitative multiple case study design, data were collected from 68 participants through interviews, focus groups, open-ended questionnaires and document reviews. Thematic analysis revealed that students integrate AI tools as cognitive extensions, enhancing understanding, argumentation and metacognitive reflection, while lecturers noted benefits alongside challenges in evaluating independent learning. Emerging approaches including oral presentations, reflective journals, project-based tasks and AI-transparent reporting were identified as effective. The findings underscore the need for AI-inclusive assessment frameworks emphasizing process, reflection and critical engagement.