Imagine a dedicated science teacher in a classroom in Dhaka, posing a thoughtful question to his students. "Think about this overnight," he says. "Whoever can reason through it tomorrow will have achieved something truly their own."

It is a noble gesture — a small ritual of intellectual cultivation designed to challenge young minds, connect them, and ultimately help them understand the world around them.

In a different world, the next day would see a handful of students with tentative answers — the fruits of their mental labour. But in our world, tomorrow promises a very different outcome. The teacher will be met not with the hesitant hands of thinkers, but with a chorus of perfectly recited, analytically sound answers.

For the night, the challenge was not wrestled with in the quiet of a student's mind; it was delegated. The question was fed into an AI model, which promptly returned the cognitive labour — pre-packaged and flawless — bypassing the very process the teacher hoped to inspire.

The sacred contract between teacher and student — the promise that effort leads to understanding — is being broken not by malice, but by a tool of irresistible convenience. The very exercises meant to build strong, independent minds are now being completed by an artificial brain.

There is no exact data on how many students from post-primary to university level (class 6 to 12) are using AI. 

According to Education Statistics 2023, the total number of students in post-primary education, including secondary school education (8,166,188), school and college (1,609,738), and college education (4,631,126), amounts to roughly 1.44 crore, excluding English-medium institutions.

Approximately 50% of the national population has internet access; however, not all students can utilise it effectively. If we conservatively assume that around 40% of these students have some level of access to AI tools, that gives us an approximate figure of 50-60 lakh students who could already be using AI for their studies.

This leads to an apparent inequality. Students in urban areas with internet-enabled devices can rely on AI for quick answers, while a large proportion, perhaps 70%, still flip through guidebooks or struggle to make sense of creative questions without such assistance. In other words, AI has quietly created a new form of educational privilege.

When the creative question system was introduced in Bangladeshi schools around 2010, it was hailed as a revolutionary step. The idea was simple but powerful: move away from rote memorisation and encourage students to think critically.

It was designed to nurture reasoning, analysis, and imagination. And for some years, teachers reported that students became more curious, asking "why" rather than just "what." That assumption is now not just outdated; it is a fantasy. The genie of Generative AI is out of the bottle, and it answers "why" faster than any student ever could. 

For students, AI feels like magic. A complex math problem? Photomath solves it with a step-by-step breakdown in seconds. An essay on climate change? Written in perfect English by Claude.ai. Even Bangla literature's creative questions can be answered instantly. 

But this convenience is a trap. We are witnessing the great outsourcing of the thinking process itself. The student is no longer a thinker but a prompt manager, a middleman between the exam paper and the AI. The "creative" in the creative question now risks standing for "creatively prompting an AI." The illusion of learning is complete when the student believes the answer is theirs, even though the process of arriving at it was entirely external.

The bigger issue is that our exam system assumes only the student's brain is available during preparation. That assumption is now outdated. Our learning and exam system is designed for a pre-AI world. Creative questions no longer guarantee creative answers. AI generates them instantly, often surpassing the work of most students under exam pressure.

This creates an apparent mismatch. Exams are meant to measure a student's critical thinking skills, but they often end up reflecting who has access to the internet or who can craft the most effective AI prompt. Education risks shifting from personal growth to digital shortcuts. 

Of course, some might argue that students have always had access to guidebooks or private tutors. But even when consulting multiple books, notes, or a tutor, the student's brain was actively working through the problem. Today, with AI, that cognitive effort is partially or even entirely outsourced, leaving little room for genuine intellectual development.

Using AI detection tools or forbidding students from using AI in education is no longer a viable solution. The real challenge lies in adapting our teaching and assessment systems to meet the demands of this new reality. In Bangladesh, this means rethinking how we teach and evaluate students, and how we prepare both learners and educators for an AI-integrated future.

Examinations must evolve to focus on skills that emphasise reasoning, analysis, and creativity, qualities that AI cannot easily replicate. Instead of relying on rote memorisation or predictable written answers, assessments should include oral exams, spontaneous essays written in class, and collaborative presentations. These formats make it harder to rely solely on AI-generated content and encourage genuine understanding, reflection, and intellectual ownership. 

AI / Artificial Intelligence / Education