– Aishwarya Rao, Director, Vivekalaya Group of Institutions

Schools across India have embraced AI with remarkable speed, and on the surface, the transformation looks real. Projectors now run algorithms instead of displaying static slides. Tablets deliver what feels like personalised content. Auto-graders process student submissions instantly without human delay. Administrative dashboards track student data in granular, real-time detail. Walk into these classrooms, and you will see technology everywhere, doing what schools promise it should do.

But dig deeper, and something becomes uncomfortable. Most AI-enabled classrooms have simply replaced traditional tools with digital versions of the same thing. A projector became an AI projector. A worksheet became an interactive worksheet. A grade book became a predictive analytics dashboard. The technology absolutely changed. The way students actually learn, the way teachers actually teach, the fundamental assumptions about what education is for—those remained untouched. Pedagogy did not shift. Thinking did not deepen. This is not integration. This is an expensive, sophisticated performance.

Where AI actually works

Real AI integration happens when schools are willing to redesign what and how they teach, not simply what device delivers it. Adaptive assessments offer a useful example of this distinction. Traditional tests measure what students know on a single day under specific conditions. Adaptive assessments do something different—they reveal patterns. They show how a student struggles with particular types of concepts, where explanations actually help them understand, where they need support before they visibly fail and disengage.

Teachers miss these patterns all the time, caught between thirty students and the curriculum they need to cover. AI does not miss them. When teachers use this data to adjust instruction in real time, something shifts. Learning changes. Not because the technology is flashy, but because teaching finally becomes responsive to actual student needs.

Personalised learning pace is another genuine application worth understanding. Different students grasp concepts at genuinely different speeds—this is not a controversial idea, yet traditional classrooms ignore it completely. Teachers move through the curriculum on a predetermined schedule, regardless of who is actually keeping up and who is just memorising to survive the test. AI-enabled systems allow students to progress at their actual pace while still maintaining access to teacher guidance and community. A student wrestling with fractions does not get left behind because the class moved to decimals. A student ready for algebra does not waste cognitive energy reviewing concepts they already understand. The pace adapts to the learner, not the learner to the pace.

These applications share something fundamental in common: they solve real, everyday problems that teachers face. They create genuine space for individualisation, not the theatrical performance of it.

Where it falls apart

Chatbots in classrooms are a case study in what performative AI looks like. Schools introduce chatbots promising students instant help with homework questions and conceptual confusion. The technology works exactly as promised. The problem is that students ignore them almost entirely. A student stuck on a problem wants an explanation from someone who knows them, understands their particular learning style, and can sense the difference between genuine confusion and distraction. A chatbot generates plausible answers quickly. That operational speed is not the same thing as teaching, and students know the difference.

Auto-grading systems reveal another critical gap in the logic. They score answers with speed that humans cannot match, which seems valuable until you realise they grade based on keywords and pattern-matching. A student who genuinely understands a concept but phrases the answer differently gets marked wrong. A student who memorised an explanation verbatim gets full marks. The system cannot distinguish between real understanding and clever retrieval. Teachers can do this, at least when they have time to actually read student work instead of managing administrative overhead.

“AI curriculum” marketed as innovation often amounts to repackaged worksheets delivered through modern interfaces. The format is digital. The presentation is sleek. The pedagogy is unchanged. Students still fill in blanks, select multiple-choice answers from scaffolded options, and move to the next module. The content simply runs on a server instead of paper, which looks innovative until you realise nothing about how students think or learn has changed.

The real question schools are avoiding

If ChatGPT can write the essay, why assign essays? If AI can solve the math problem, why teach problem-solving? If technology can deliver the answer, what is school actually for? Most schools have not asked this question seriously. They integrated AI without redesigning the curriculum or assessment. They bought tools without training teachers on when and why to use them, creating expensive, unused technology that gathers dust alongside the enthusiasm that originally greeted it.

Schools that actually did the harder work asked different questions first. What should students learn that AI cannot do? What kinds of thinking matter more than information retrieval? What human capabilities grow more valuable precisely as technology handles routine tasks? Only after answering these questions did they consider where AI could genuinely support their vision.

The integration gap

Real AI integration requires teacher training that goes far beyond platform tutorials. Teachers need to understand what problems AI actually solves and what problems it creates. They need to redesign what happens in classrooms if they are using AI for assessment and pacing. They need to shift fundamentally from delivering information to designing experiences where students think, struggle, and genuinely grow.

This is harder than purchasing technology. It requires time, resources, and willingness to disrupt systems that currently work. Many schools will skip this step. They will have the tools without having thought through what to do with them.

Schools that do the harder work will look visibly different. Classrooms will be quieter because students will be thinking more than listening. Teachers will talk less and ask more questions. Assessment will measure actual capability rather than recall. Technology will serve a clear purpose rather than create novelty for its own sake.

That is the actual difference between having AI in classrooms and having genuinely AI-enabled classrooms.

Also read: Assessing AI skills: What can ‘good’ AI questions tell us about how students think?