Related Experiment Videos
Designing AI-resilient assessment in higher education: a four-pillar conceptual framework
Dragan Nikolić1, Marijana Basta Nikolić1
1Faculty of Medicine, University of Novi Sad, Novi Sad, Serbia.
Frontiers in Artificial Intelligence
|July 31, 2026
Summary
Generative AI challenges traditional academic assessments. This paper proposes an AI-resilient framework focusing on demonstrable reasoning and learning ownership, not just written output.
Area of Science:
- Educational Technology
- Academic Integrity
- Assessment Design
Background:
- Generative AI (AI) tools can generate academic text, challenging the validity of traditional assessments.
- Current AI detection methods have limitations in accuracy and equity.
- Rethinking assessment is crucial in the age of AI.
Purpose of the Study:
- To propose a framework for AI-resilient assessment.
- To shift evaluation from product quality to demonstrable learning processes.
- To offer practical tools for implementing AI-resilient assessments.
Main Methods:
- Framework development based on four pillars: process-based documentation, oral defense, authentic tasks, and transparent AI-use policies.
- Operationalization through a rubric model, oral-defense protocol, and assessment vulnerability audit tool.
- Focus on design instruments for future empirical evaluation.
Main Results:
- The paper presents a conceptual framework, not empirical data.
- Proposed tools are design instruments for future validation.
- The framework is illustrated using health sciences education but applicable to other disciplines.
Conclusions:
- Generative AI necessitates a paradigm shift in assessment strategies.
- AI-resilient assessment should prioritize evidence of reasoning, decision-making, and learning ownership.
- The proposed framework offers a proactive approach to integrating AI in education.
