Rethinking Assessment in the AI Era: Emerging Challenges and Responses

Abstract

The widespread adoption of generative artificial intelligence (AI) has prompted renewed attention to how academic writing should be assessed in higher education. While AI supports multilingual writers throughout the writing process, it also challenges traditional product-based assessment by obscuring students' learning, authorship, and decision-making. Guided by Sociocultural Theory, this qualitative descriptive study explored how assessment can be redesigned for AI-mediated academic writing. Data included classroom reflections, student writing artifacts, multiple drafts, revision records, and AI-supported writing activities. Data was analyzed using Braun and Clarke’s (2023) Reflexive Thematic Analysis. Thematic analysis generated six interconnected themes that reconceptualize assessment as the documentation of learning rather than the evaluation of final products alone. The findings suggest that process-oriented assessment, including scaffolded drafting, revision histories, reflective writing, AI-use disclosures, and ongoing feedback, provides richer evidence of students' cognitive engagement, ethical source use, and writing development. The study proposes a process-oriented assessment framework that promotes transparency, responsible AI use, learner accountability, and more authentic assessment in multilingual writing classrooms.