From prompting to thinking: Redefining academic integrity in the era of AI — Nor Farah Hanis Zainun

On September 12, the reality for academics has become increasingly familiar: students are turning to artificial intelligence (AI) tools like ChatGPT and Gemini for their academic tasks. AI is no longer something universities can keep outside the classroom; it has become embedded in the digital environment where undergraduate assignments, postgraduate research, and everyday learning now take place.

Rather than merely restricting or detecting AI use, universities should focus on redefining academic integrity by considering what exactly they are trying to assess. A major flaw in current thinking is assuming students can be separated from technology; today's students have grown up with smartphones and instant information access, making technology a fundamental part of their learning environment.

The challenge is not that students are using AI tools but that many assessment models still heavily rely on the final product, such as polished reports or essays, while offering little insight into the human process behind it. If students can convincingly complete an assignment with a single AI prompt, perhaps the assessment should question what the assignment was designed to measure.

Protecting academic integrity requires rethinking how evidence of learning should be presented. Rather than focusing solely on detecting or restricting AI use, universities should redesign assessments to evaluate students' critical thinking and intellectual contributions when using AI. Simple measures like requiring a concise AI-use log or a selected prompt trail could make the student's intellectual contribution visible, such as identifying key prompts used, highlighting accepted or rejected suggestions, pointing out factual errors, and explaining how they verified information using credible sources.

Students could also interrogate AI-generated answers, comparing different responses, testing arguments against academic literature, or explaining why seemingly convincing answers are actually weak. This shifts the student's role from a passive recipient to a critical evaluator. Redesigning assessments to include multiple stages of the student's intellectual journey, from initial idea to final revision, can also help.

While academic misconduct remains a concern, AI raises new challenges; academic integrity should become more demanding, expecting students to demonstrate judgment, verification, intellectual ownership, and the ability to defend their choices. The question should not only be "Did the student use AI?" but "What intellectual work did the student still have to do?" Redesigning assessments around this question could redefine academic integrity in the age of AI.