An IIT Kharagpur professor has sparked a fresh debate on technology in education with a blunt warning: when artificial intelligence does students’ academic work for them, there is no real learning taking place. The argument cuts to the heart of one of the biggest questions facing universities today — how to keep AI as a tool without letting it hollow out education.

The professor’s point is deceptively simple. Learning is a process: struggling with a problem, making mistakes, testing ideas and gradually building understanding. When an AI system produces the finished answer at the click of a button, the student skips every one of those steps. The assignment gets submitted, the grade may look fine — but the understanding that the assignment was supposed to build never forms.

The classroom dilemma

This is not an abstract concern. Across Indian campuses, educators report that students are increasingly turning to generative AI for essays, problem sets and even coding assignments. The result is a growing gap between what grades suggest and what students can actually do — a gap that shows up in examinations, interviews and workplaces.

The challenge for institutions is that AI is not going away, and banning it outright is both impractical and arguably counterproductive. The technology is already part of the professional world students are training to enter. The question is where the line falls between using AI to learn faster and using it to avoid learning at all.

What experts are suggesting

Academics wrestling with this problem are converging on a few approaches. Assignments are being redesigned to require original thinking that cannot be outsourced — in-class work, oral defences, process-based evaluation and tasks tied to the student’s own context. Some institutions are teaching AI literacy explicitly, so students learn to use the tools critically rather than uncritically.

The underlying principle, echoing the professor’s warning, is that education must protect the process, not just the product. A submitted assignment is evidence of work done — not proof of learning achieved. Until assessment catches up with the technology, the risk is a generation of graduates with polished credentials and missing foundations.