AI-Aligned Assessment Practices for a Future-Ready Higher Education System

Abstract

The rapid integration of generative artificial intelligence (GenAI) into higher education has prompted institutions to reconsider traditional assessment methods. As students increasingly use GenAI for brainstorming, drafting, and problem-solving, assessments based on memorisation or predictable outcomes no longer reliably measure independent learning or authentic skill development. Experimental evidence confirms that AI-generated submissions can achieve pass-level grades and that markers cannot reliably distinguish AI-authored work from student work, even in tasks designed to be authentic. This study conducts a systematic review of 25 peer-reviewed articles (2022–2026) examining generative AI, assessment, and academic integrity in higher education, complemented by practice-based illustrations from Commerce modules at a South African private higher education institution. Addressing the shortage of integrated, task-level guidance for assessment redesign, the review proposes the AI-Aligned Assessment Framework: five interconnected principles—Critical Judgement, Ethical AI Use, Scenario-Based Application, Learning Process Visibility, and Transparent Student–AI Collaboration. The framework treats AI capability as a moving target rather than a fixed boundary and is designed to be scalable across disciplines and institutional contexts, including those characterised by high student-to-staff ratios. Rather than designing primarily to detect or exclude AI use, the framework shifts assessment design towards generating credible evidence of student reasoning, accountability, and contextual application. The paper contributes a practical, discipline-adaptable framework that supports academic integrity, authentic learning, and graduate preparedness in AI-enhanced higher education.