A silent student has entered every classroom. It does not wait outside the school gate, raise its hand, borrow a book, or sit for an examination. It lives in the mobile phone, the search bar and the laptop. It can translate a paragraph, solve a mathematical problem, write an assignment, and even produce a research proposal in seconds. Its name is artificial intelligence (AI).
For Bangladeshi learners, AI is becoming unavoidable for practical reasons. Students use it to understand difficult English texts, practice pronunciation, summarise lectures, and receive instant feedback on writing. Teachers can create worksheets and adapt materials for mixed-ability classes. Universities can support research, administration and career preparation. Where large classes, uneven English proficiency and limited teacher time remain persistent challenges, AI can offer individual support.
But convenience is not the same as learning. A student who asks AI to write every paragraph may submit fluent English without developing the ability to think, argue, or write independently. A researcher who accepts generated references without checking them may spread false information. A teacher who uses an AI-generated lesson without reviewing it may bring bias, error, or cultural irrelevance into the classroom. AI can make weak work look polished. That is precisely why it is both powerful and dangerous in our educational institutions.
An assignment may appear original while the student has not understood a single sentence. Unequal access may widen inequality: affluent students can use paid tools, better devices, and faster internet, while disadvantaged learners may be left behind. AI systems can provide inaccurate information with great confidence. Students may become dependent on instant answers and lose patience for reading, drafting, revising, and struggling with a problem. AI tools reflect the languages and assumptions embedded in their development. If we use them carelessly, our local knowledge, Bengali expression, and regional realities may become invisible. Grammatically correct English does not guarantee sensitivity to the social texture of a Bangladeshi classroom or the cultural meaning carried by our languages.
Developed education systems are establishing guardrails for AI. Australia’s framework for schools addresses fairness, accountability, and privacy; Singapore provides AI literacy modules for students. These approaches offer Bangladesh useful guidance: responsible use and protection of learners. Our response must begin with a needs analysis.
AI should be used because it solves a real educational problem. In English language teaching, it may help learners practice conversation and explore discipline-specific vocabulary. In ESP courses, it can simulate workplace emails, job interviews and technical communication. But it should not replace the teacher’s judgment, the learner’s voice or genuine interaction.
Our teaching materials, methods, aims and objectives must therefore be redesigned. Textbooks should include tasks that require students to compare AI-generated answers with reliable sources, identify bias, correct errors, and explain their reasoning. Learning objectives should move beyond memorising information towards evaluating information, asking good questions, collaborating responsibly and applying knowledge in real contexts. Schools and universities should include process portfolios, oral presentations, classroom demonstrations, reflective journals and short viva examinations. Students should explain how they reached their answers. An AI-use declaration stating the tool, its purpose, and how the output was checked can make academic honesty more transparent.
Teachers stand at the centre of this transformation. They cannot address AI through a one-day workshop or a vague circular. Teacher development must include practical AI skills, ethical use, privacy protection, bias detection, source verification and assessment redesign. Teachers need time, institutional support, and professional confidence. A prepared teacher will know when AI can assist learning and when students must work without it.
AI literacy should be integrated gradually from primary education to postgraduate study. Primary children can learn digital safety and fact-checking through age-appropriate activities under careful supervision. Secondary students should learn how algorithms shape information and how to verify claims. At university, AI should be connected to disciplinary learning: future engineers, doctors, teachers and journalists need to understand how it affects their professions. At the postgraduate and doctoral levels, AI can assist with literature mapping, language editing, coding support, and research organisation. But researchers must never present invented data or false citations as evidence. Research originality means that intellectual responsibility remains human. Universities should require transparent disclosure, maintain version histories, strengthen supervision, and use oral defence to ensure that researchers can explain their claims.
We should also invest in Bengali and other languages. AI can help preserve local stories, build digital dictionaries, improve translation and make educational content accessible. Bangladeshi educators, researchers, and communities must help create reliable local-language resources and review their accuracy.
Our task is more demanding: to make AI our informed companion. We must teach students to use it with curiosity, caution, creativity and conscience. The future will belong not to those who can merely ask a machine for answers, but to those who can question the machine, improve its answers and remain fully human while using it.
The views expressed in this article are solely those of the author
The writer is an Assistant Professor of English, IUBAT, and a DRS at UPM, Malaysia
