India's classrooms in the age of AI: where we stand, where we're headed, and what it means for jobs

Built for exams, not for thinking

India has built one of the largest school systems on earth, but it was designed to produce clerks and exam-toppers, not thinkers. The next ten years of AI will expose that gap faster than any reform committee could.

The numbers frame the stakes. About 37 crore Indians are aged 15 to 29, more than a quarter of the population. Nearly every child is now enrolled in school, and more young people than ever hold degrees. Yet most children in Class 3 cannot read a Class 2 textbook, and close to 40% of graduates under 25 who are looking for work cannot find it.

Now add AI: tools that can already write, code, translate, tutor and summarise better than many entry-level employees. This post looks at three questions: where Indian education stands today, how AI will reshape it by 2036, and what India must do so that its largest generation ever is not also its most frustrated one.

Where Indian education stands today

India has solved access but not learning. In rural India, 98.1% of children aged 6–14 are enrolled in school, yet most leave without the basic skills the system promises.

Foundational learning is weak

The ASER 2024 survey tested children in rural districts across the country. Only 23.4% of Class 3 children in government schools could read a Class 2 text, up from 16.3% in 2022. The recovery after Covid is real, but three in four still can't read at a level two years below their own, and learning levels vary sharply from state to state.

Infrastructure is improving, unevenly

UDISE+ 2025-26 counts about 14.67 lakh schools, 1.03 crore teachers and 24.72 crore students. Internet now reaches 67.4% of schools, up from 49.7% three years earlier, and 95% have electricity. That still leaves roughly 4.8 lakh schools offline and about 73,000 without power.

Rote learning and the coaching economy

The system still rewards memorising for a few high-stakes exams: board marks, JEE, NEET, CUET and government job tests. Families spend heavily on coaching, and repeated paper leaks have shaken trust in those exams. Students learn to answer questions, not to ask them.

Degrees without jobs

Higher education has grown fast: the gross enrolment ratio for ages 18–23 reached 30% in 2023-24. But Azim Premji University's State of Working India 2026 found that close to 40% of graduates under 25 who are looking for work are unemployed, and unemployment among young graduates has stayed between 35% and 40% since 1983. In 2023, about 1.1 crore of India's 6.3 crore graduates aged 20–29 were unemployed.

Even finding work doesn't mean a good job. Among young male graduates who were unemployed, about half found work within a year, but only around 7% found a permanent salaried job. Meanwhile, the India Skills Report 2026 rates 56.35% of graduates as employable.

The reform is real, but slow

The National Education Policy (NEP) 2020 points in the right direction: mother-tongue teaching in the early years, foundational literacy missions, flexible degrees with multiple entry and exit, vocational exposure from Class 6, and less weight on rote exams. From the 2026-27 session, CBSE schools teach computational thinking and AI from Class 3 to Class 8. The problem is the gap between policy on paper and what happens in a rural classroom with one teacher and no internet.

How AI will reshape Indian education by 2036

By 2036, AI will not replace schools, but it will replace the parts of schooling that are about transmitting information. What remains will be the human parts: motivation, judgement, collaboration and character.

1. A personal tutor for every child, in her own language

The biggest promise for India is the one-to-one tutor that only rich families could afford. AI tutors that speak Hindi, Tamil, Bengali or Marathi, adjust to each child's pace, and explain a concept five different ways can attack the foundational learning crisis directly. Indian-language models are already being built under the IndiaAI Mission, including BharatGen's Param2, which covers 22 Indian languages.

2. Teachers become coaches, not lecturers

In a single-teacher school with children across three grades, AI can take over drills, worksheets, translation and grading. The teacher's job shifts to spotting who is stuck, mentoring, managing the classroom and building values. This only works if India invests in teacher training at scale: it has 1.03 crore school teachers, and most of them will need it.

3. The end of the rote exam

When any student can ask an AI to write an essay or solve a physics problem, exams that test recall lose meaning. Expect a shift to projects, oral defence, open-book problem solving, portfolios and continuous assessment. Boards and entrance exams that do not adapt will measure who has the better chatbot, not the better mind.

4. Skills over degrees

Employers will care less about the name of a degree and more about proof of ability: what you built, what you shipped, what you can demonstrate. Short, stackable credentials, apprenticeships and the Academic Bank of Credits under NEP fit this future better than a fixed three-year course.

5. Learning never stops

If skills go stale every few years, education cannot end at 22. AI-guided upskilling at 30, 40 and 50 will become as normal as college is today.

The risk is just as real. AI can also make students lazy thinkers who outsource every effort. And if only private, urban schools get good devices and trained teachers, AI will widen the gap between India's two school systems instead of closing it.

Where India will stand in 2036

India can become the world's largest AI-skilled workforce, or the world's largest pool of educated, unemployed youth. Both outcomes are on the table, and the next five years decide which.

The likely picture

If India fixes the basics and keeps its momentum, by 2036 it could be the world's main supplier of people who apply AI: building products for Indian languages, running AI systems for global firms, and serving the next billion users. If it does not, AI will make the top 10% of Indian students world-class while leaving the rest further behind.

Tackling the jobs question

AI is not wiping out jobs in India yet, but it is closing the front door for freshers. That is the problem to solve.

What the data shows so far

• NASSCOM estimates the tech sector added a net 1.35 lakh jobs in FY26, barely more than the 1.33 lakh in FY25, for a total of about 59.5 lakh.
• TCS said in April 2026 it had made 25,000 fresher offers for FY27, with more depending on demand; reports put its FY26 fresher hiring at about 44,000.
• Demand for AI skills keeps rising: Nasscom and Deloitte expect India's AI talent need to roughly double from 6–6.5 lakh to about 12.5 lakh professionals by 2027.

In short, it's a two-tier job market. Experienced people who use AI well are in demand. Freshers who can only do routine coding, testing, data entry or basic customer support are being squeezed.

What the government should do

1. 1Fix foundations first. Make reading and arithmetic by Class 3 a non-negotiable national target, tracked by district.
2. 2Reform assessment. Move board and entrance exams away from recall towards reasoning, projects and applied problems.
3. 3Make apprenticeships the norm. Tie every degree to real work experience, with tax incentives for firms that hire and train freshers instead of only lateral hires.
4. 4Fund lifelong reskilling. Give every adult a skilling account, building on Skill India, to use at any age for certified AI-era courses.
5. 5Back labour-intensive sectors. AI can't be India's only jobs plan: manufacturing, tourism, healthcare, construction and the care economy must absorb millions.
6. 6Build public AI for Bharat. Use Indian-language AI to bring tutoring, legal help and farm advice to small towns, creating local jobs in deployment and support.

What students and parents should do

• Learn to think, not to memorise. Reasoning, clear writing and problem-solving will outlast any syllabus.
• Use AI as a sparring partner, not a crutch. Ask it to quiz you and challenge your answers, not to do your homework.
• Build a portfolio. Projects, internships, open-source work, a small business: proof of ability beats marks.
• Pair a domain with AI. AI plus law, AI plus agriculture, AI plus healthcare is far stronger than generic "AI skills".
• Invest in human skills. Communication, teamwork, empathy, negotiation and leadership are the hardest to automate.
• Don't chase only government jobs or a single entrance exam. Keep options open.

What working professionals should do

• Use AI tools daily in your current job, so you become the person who is three times faster, not the person who is replaced.
• Move up the value chain: from doing tasks to designing, reviewing and owning outcomes.
• Set aside time every year to learn a new, adjacent skill.

The bottom line

A country that teaches its children to read, reason and keep learning will turn AI into the greatest equaliser in its history. A country that keeps rewarding memorisation will watch AI do the memorising better, and its youth will pay the price.

The tools are arriving either way. The choice in front of policymakers, schools, parents and students is simple to state and hard to execute: stop preparing young Indians for the exams of the past, and start preparing them for the work of the next decade.