In the evolving landscape of higher education, educators are reconsidering how students engage with artificial intelligence (AI) tools, emphasizing critical thinking over mere content production. Dr. P. Rajesh Kumar, deputy director at the Centre for Research & Innovation at Open University Malaysia, advocates for a shift in teaching strategies to better prepare students for the widespread integration of AI in academic work.
Dr. Kumar argues that if AI chatbots can generate acceptable assignments within seconds, educational institutions must question the original purpose and design of these assessments. Instead of outright banning AI, he proposes rethinking how students are taught and evaluated from the outset of their studies.
While traditional proctored examinations, whether in controlled physical or online environments, remain relevant, Dr. Kumar suggests incorporating AI-generated responses into the learning process in new ways. He recommends that students be presented with AI-produced answers and tasked with critically analyzing them. Drawing on the investigative style of the fictional detective Columbo—known for his persistent questioning and refusal to accept evidence at face value—Dr. Kumar encourages students to adopt a similar approach toward AI-generated content.
“AI literacy for students should begin from Day One with one simple non-negotiable rule: never take AI output at face value,” he wrote. Students must learn to interrogate AI responses by verifying sources, assessing the validity of arguments, identifying assumptions, and recognizing omissions.
Further, Dr. Kumar envisions exercises where students use one AI system to generate arguments and a second to challenge or critique those arguments, requiring students to evaluate and synthesize the conflicting information. This method, he contends, fosters deeper analytical skills compared to traditional assignments that focus on content creation alone.
He extends this concept to examinations, proposing that students be given complex scenarios accompanied by AI-generated solutions. Their task would be to detect errors, question assumptions, explore alternative viewpoints, and ultimately justify their own conclusions.
Highlighting the rapid pace at which AI can produce information, Dr. Kumar underscores that the essential competency is not generating answers swiftly but discerning when those answers are incorrect, incomplete, or misleadingly persuasive.
He cautions that if education continues to prioritize content production without integrating critical evaluation skills, there is a risk of training students to replicate tasks machines can already perform more efficiently. Instead, universities should focus on nurturing judgment and skepticism, echoing Columbo’s hallmark phrase: “Just one more thing.”