top of page

AI and Education: If Information Is Everywhere, What Is Left to Learn?

Writer: Dr Sp Mishra
Dr Sp Mishra
Oct 1
8 min read
AI and Education
AI and Education

AI has made information nearly free. The harder question is what education should do now.


A student who misses a point in class can now ask an AI tool to explain it again, simplify it, translate it, or illustrate it, all within seconds. Learning has become more interactive and more personalised than it has ever been.


But easier access to information does not mean more learning. That was the thread running through my recent conversation on the India Career Centre Podcast with Dr Aruna Parandhama, who teaches English at an institution in Bangalore. She began her career at Goldman Sachs and moved into teaching in 2013. Since then she has worked across secondary, higher secondary and higher education, and completed a PhD along the way. I first came across her through her article "Don't Teach, Steer" in The Hindu, and our conversation grew out of it.


Explanation is not understanding


Even the youngest students now know that information is a click away. As Dr Aruna put it, that is empowering, and it is also a double-edged sword. A student who knows they can catch up later has less reason to listen today.


Her test is simple. Whatever the medium, ask whether comprehension is still happening. Receiving an explanation and understanding something are different things, and a clear AI answer can hide the gap between them.


Some difficulty is worth keeping


I asked whether we risk losing the struggle of learning: rereading a paragraph, staying with an unfamiliar idea until it makes sense. Dr Aruna was not afraid of this, but she was clear that some struggle is necessary.


She referred to "desirable difficulties" from educational psychology, a term coined by the cognitive psychologist Robert Bjork. The idea is that difficulty, deliberately placed in a learning journey, makes the brain process material from several angles and retain it longer. Much of the supporting evidence comes from laboratory and classroom studies that predate generative AI, so how well it carries over to learners with an AI tutor on hand is still an open question.


Her practical example was a lesson built in layers. AI can help with the early stages: a video, an animation, worked examples. Those stages are about remembering and understanding, and making them easier does no harm. The difficulty belongs higher up, where students must apply what they have learned to a situation they have not seen before. Deciding where the struggle goes, and how much of it, is the teacher's skill.


The classroom has to change with AI


Dr Aruna described how her own teaching has changed. She once relied on sustained discussion and long stretches of reading. Now, with early-morning classes of around 80 students and shrinking attention spans, she breaks each session into segments: a puzzle or image on the board, a provocative line to contest, ten minutes of student responses, then back to reading, then another break. She was candid that this has made teaching more performative than it used to be, and equally candid that she sees no point resisting it.


Her most striking example was a class on B.R. Ambedkar. Her students came largely from privileged business families and knew the name without feeling its weight. Instead of telling them he was a great man, she asked a student who needed to make up a missed assessment to play Ambedkar. He had to read up on both the achievements and the failures, and defend himself while the class put him on trial. Anyone who praised Ambedkar was challenged by the student playing him, and anyone who criticised him had to research their points first. Each side also had to be able to argue the opposite position. She called the activity "sink or float" and played only the role of fact-checker. The class, she said, expected something heavily political and got a balanced, productive discussion instead.


Her takeaway was that students are a teacher's greatest resource, and that educators must be willing to walk further than halfway towards them.


AI's Impact: Is Deep reading and writing at risk


Dr Aruna was sympathetic to young people here. They are born into systems that reward speed and novelty and deliver information in bite-sized pieces, so sustained reading looks like a dying habit. Writing faces the same pressure. Her sharpest observation was about communication shrinking from the carefully composed message to the SMS to the emoji, and now to gestures that stand in for whole sentences.


Reading and writing are not just ways of moving information around. They are how we build and organise thought, and that matters more, not less, when AI can produce polished text on demand. Students need to learn how to use AI, and also when not to.


Access is not the real divide


This was where the two of us saw things slightly differently, and the difference is worth keeping.


I argued that AI could help break the asymmetry of information. Even at the free tier, a student without a good teacher, library or coaching class can now get explanations that were once out of reach. Dr Aruna agreed, but argued that in India, getting hold of a device is no longer the main constraint.


The ASER 2024 survey of rural India supports part of this. Nearly 90% of 14 to 16 year olds report a smartphone at home, and 82% know how to use one. But the same survey shows where the gap sits. Only about 31% of these teenagers own a phone themselves, with 36% of boys against 27% of girls. And while 82% can use a smartphone, only 57% report using it for an educational activity, compared with 76% for social media. Access to a device is close to universal. Using it to learn is not.


That fits her argument. A student who can ask a question but cannot apply the answer to a different problem is stuck, however many subscriptions they hold. Most students, she noted, have no orientation in AI ethics and still use these tools quietly.


My own view is that both things are true. AI can narrow the information gap and widen the capability gap at the same time. Which effect dominates will depend on whether schools teach students to question AI, verify it and use it responsibly.


Assessment must look beyond the final answer


Traditional assessment evaluates the product of learning. When AI can produce the product, it says little about the learner. Dr Aruna's alternative is to assess the journey. For an undergraduate research project, that means asking what the student did not know at the start, how they learned about it, what problems they met, what extra reading they did, and how their final view differs from their first.


She drew on a visit to Monash University and the University of Melbourne in Australia, where faculty and students told her AI is the way to go. She added that in some select universities, students are asked to submit the prompts they used, which shows how they engage with the system and not just what it produced. India is not at that point yet, she said, but she saw the logic of it. She was also honest about the limits. A viva can cover part of this but is not foolproof, and more research is needed.


Her preference is for small-stakes assessments that test application: one based on classroom discussion, one on a practical experiment, open-book formats that ask students how they will build on what they have read, and no more purely recall-based questions.


Then she said the thing every Indian educator will recognise: "our numbers are a big challenge." Process-based assessment is far easier with small classes. With 80 students in a room and a large national examination structure above it, it is a serious constraint. I suggested that universities need more autonomy to experiment, and she agreed that small changes made consistently can change the path.


Feedback is where teachers become irreplaceable


Dr Aruna's strongest argument came when I asked what she would change if she could redesign the system from scratch. Her answer was feedback. Not a mark of 64 out of 80, but real, individual, humane feedback from a teacher who tells the student: I am interested in journeying with you.


She uses AI for this herself. During a presentation she jots short notes on gesture, language and confidence, then uses AI to turn them into a well-written paragraph for each student. AI handles the writing, while the observation and judgment stay with the teacher. She also uses it for clerical work and for finding references. This is her line between low-value and high-value tasks. Few institutions make time for individual feedback, she said, and that is a disservice to the profession.


If machines can explain almost anything, this is what is left for the teacher: to be a guide who helps students make sense of what they have been given.


What students should keep building


I asked what students should focus on when knowledge is available everywhere. Her answer fell into four areas.


  • Foundational knowledge from sustained, analog sources. Books, newspapers and autobiographies, in preference to hours of clips. She reminded me that oral traditions survived for centuries because they rest on memory.

  • Organic communication. Write and speak in full sentences, not only symbols.

  • The habit of contesting. Question claims, triangulate information, read up further, and do not accept anything without challenging it.

  • Knowing yourself as a learner. When and how you learn best, and what works for you.


Not humans versus AI


The useful question is not whether humans or machines will win. AI can make information more accessible, explanations more personal and routine tasks lighter. Education's job is to keep the learner actively involved.


If students use AI to avoid thinking, learning weakens. If they use it to ask better questions, test their understanding and explore unfamiliar ideas, it strengthens. Dr Aruna's own rule is simple: everything should be an exercise in thinking.


The future classroom may not need teachers who know everything. It will need teachers who can help students make sense of almost anything. And when information is abundant, education can stop asking how much a student can remember, and ask whether they can understand, question, communicate and apply.


https://youtu.be/5CVuhtem2d0
My Conversation with Dr Aruna Parandhama

Continue the conversation


This episode is part of a larger set of conversations on the India Career Centre Podcast about learning and the future of higher education:



The question of education in the AI era did not begin with generative AI, and it will not end with it. AI simply makes it more urgent.



References



Comments

bottom of page