Indian Institute of Technology Kharagpur (IIT Kharagpur) professor Plaban Kumar Bhowmick, from the Department of Artificial Intelligence and principal investigator of the National Digital Library of India, has highlighted a crucial distinction between cognitive offloading and cognitive scaffolding in the use of artificial intelligence (AI) in education.
Explaining the difference between the two concepts, Professor Bhowmick said educators use cognitive scaffolding to create a supportive structure that enables students to learn independently.
"We do something that we call cognitive scaffolding, where we put up a structure within which the student feels supported. Now, AI can take a problem given by a student and solve it on the student's behalf. So, the question is whether AI in this form is actually helping the student. In my view, it is not really helping because this is what we call cognitive offloading or cognitive hijacking. Basically, there is no learning happening in this scenario," he said.
"The AI is doing the job for the student, and the learning goal is not really achieved. But if we think of AI as a form of cognitive scaffolding, the student does the entire job while the AI provides support. That is where the Fermi AI Tutor comes in. It supports students in problem-solving while they work through the problems on their own. The AI tutor supports the student but does not solve the problem for them," he added.
What if AI could help students think better, rather than simply give them answers? At IIT Kharagpur, we believe the future of AI in education lies not in replacing human thinking, but in strengthening it. Prof. Plaban Kumar Bhowmick, Department of Artificial Intelligence, IIT Kharagpur, and Principal Investigator of the @NDLIndia highlights a crucial distinction: Cognitive Offloading vs. Cognitive Scaffolding. While conventional AI tools may provide ready-made solutions, the @fermi_ai Tutor takes a different approach. It encourages students to think, explore, and solve problems independently, offering guidance and hints rather than direct answers. With students already engaging with the platform and returning to learn more, Fermi AI is demonstrating the potential of AI as a teaching companion, not merely an answer-generating tool. The future of learning isn't about AI thinking for students. It's about AI helping students think for themselves. Experience the Fermi AI Tutor, available free through NDLI: https://t.co/E6LOKs5PoU #FermiAI #NDLI #IITKharagpur #AIinEducation #FutureOfLearning @EduMinOfIndia
— IIT Kharagpur (@IITKgp) October 8, 2026
Asked whether any aspect of the Fermi AI Tutor had genuinely surprised him, Professor Bhowmick pointed to its underlying design philosophy. He said the platform could change how students perceive and use AI, moving beyond the idea of AI as a tool that users simply pick up and put down.
"What this AI tutor brings to students is a coaching companion, a teaching companion. It does not just solve problems or support students in solving them; it transforms students. This kind of AI is not merely a tool. It is becoming the future of learning," he said.
The institute said the future of AI in education lies not in replacing human thinking but in strengthening it.
"While conventional AI tools may provide ready-made solutions, the Fermi AI Tutor takes a different approach. It encourages students to think, explore and solve problems independently, offering guidance and hints rather than direct answers," the institute said.
The institute added that students are already engaging with the platform and returning to learn more, demonstrating Fermi AI's potential as a teaching companion rather than merely an answer-generating tool. The future of learning, it said, is not about AI thinking for students but about AI helping them think for themselves.
