Entrepreneurship in Action, 21-24 Nov, online. Build a working product with AI

Teaching AI in Classes 8 to 12: what a school can cover, with classroom examples

15 min read

On this page

The short answer

For 2026-27, CBSE directs schools to introduce computational thinking and AI in Classes 3 to 8 within existing periods, and offers AI as a skill subject: 417 in Classes 9 and 10, 843 in 11 and 12. A school can add practice building with AI, checking AI answers against a source and keeping personal data out of prompts. One Young India runs programmes of its own in this area, so weigh what follows with that in mind.

Facts checked on 16 November 2026 against the sources below.

How this was made: drafted with AI help from the sources listed, and checked by the One Young India editorial desk on 16 November 2026.

What does CBSE ask of a school on AI in 2026-27?

A Press Information Bureau release from the Ministry of Education (1 April 2026) says CBSE is implementing a curriculum on Computational Thinking and Artificial Intelligence (CT and AI) from Class 3 to Class 8 in the 2026-27 session.

CBSE’s Secondary Curriculum 2026-27 (Part 1) says schools are directed to introduce it with no reliance on particular platforms, devices or extra teaching time, since the competencies sit inside existing subject periods. CBSE Academic’s CT and AI page carries a Student Handbook and a Teacher Handbook for each grade from 3 to 8.

For Classes 9 to 12, Artificial Intelligence is a skill subject: code 417 in Classes 9 and 10 and code 843 in Classes 11 and 12, each with a 2026-2027 curriculum. This page keeps to AI, not CBSE’s full list of skill subjects.

Class 9 in 2026-27 has its own scheme. CBSE’s Circular No. Acad-70/2026 (29 October 2026), “Scheme of Studies for Class IX 2026-27”, applies to students in Class 9 in 2026-27, who reach Class 10 in 2027-28. For these batches, it says, the “Modules of CT & AI” in Class IX are not part of the curriculum. Class 10 in 2026-27 stays on the old scheme.

For the same students, Vocational Education “shall continue to be a compulsory curricular area”: a student takes “Kaushal Vikas”, assessed internally, or one of the 22 skill subjects the circular lists, Artificial Intelligence (417) among them, with a Board examination in that subject in Class 10 (2027-28). The circular is on CBSE Academic’s circulars list; a school planning Class 9 should read it in full.

Table 1. What CBSE’s 2026-27 documents set out, by class. CBSE’s 2026-27 documents and PIB’s release on them, read on 16 November 2026; units as CBSE names them, without the product names its syllabi cite.

ClassWhat CBSE’s document says for 2026-27What the subject-specific part coversHow it is assessed
8Part of the CT and AI curriculum for Classes 3 to 8, inside existing subject periods and low-tech; a Student Handbook and a Teacher Handbook for Grade 8Logical reasoning, pattern recognition, data interpretation, algorithmic thinking and ethical awareness, with foundational AI literacy; computational thinking first, as a basis for learning AI laterContinuous and competency-based: written tests with CT puzzles, group activities, a Teacher Observation Journal
9Artificial Intelligence, code 417, a skill subject taken along with other subjects; for the 2026-27 batch, one of the 22 skill subjects under Acad-70/2026, and the CT and AI modules are not part of that batch’s curriculumAI Reflection, Project Cycle and Ethics; Data Literacy; Math for AI (Statistics and Probability); Introduction to Generative AI; Introduction to Python100 marks (50 theory, 50 practical), 210 hours; one project, field visit or portfolio tied to the Sustainable Development Goals, 15 marks
10Artificial Intelligence, code 417; Class 10 in 2026-27 stays on the old schemeRevisiting AI Project Cycle and Ethical Frameworks for AI; Advanced Concepts of Modeling in AI; Evaluating Models; Statistical Data (assessed through practicals); Computer Vision; Natural Language Processing; Advance Python100 marks (50 theory, 50 practical), 210 hours; sample projects include predicting penguin species and sentiment analysis
11Artificial Intelligence, code 843Units include Artificial Intelligence for Everyone; Python Programming; Introduction to Capstone Project; Data Literacy; Machine Learning Algorithms; Leveraging Linguistics and Computer Science; AI Ethics and Values100 marks (50 theory, 50 practical); the capstone project carries 12 of the 50 practical marks
12Artificial Intelligence, code 843Python Programming II and a no-code data mining unit (both evaluated in practicals only); Data Science Methodology: An Analytic Approach to Capstone Project; Making Machines See; Introduction to Big Data and Data Analytics; Understanding Neural Networks; Generative AI; Data Storytelling100 marks (50 theory, 50 practical); of the 50 practical marks, capstone project with documentation and video 25, practical file 10, lab test 10, viva voce 5

What do the AI syllabi teach, and what do they not single out?

Class 9 introduces the AI project cycle, from problem scoping and data acquisition to evaluation and deployment, sets rule-based models beside learning-based ones, and names true and false positives and negatives. Class 10 adds the train-test split, accuracy, precision, recall and F1 score, with an activity to build a confusion matrix from scratch; Class 12 calculates those measures from one.

Data and ethics run alongside. Class 9 separates data privacy from data security, recommends telling a real image from an AI-generated one, and takes up bias in how data is collected; Class 11 has a unit on AI ethics and values. Building is there too, from Class 9’s suggested project, an AI model made with a ready-made tool, to designing and deploying a chatbot in Class 12.

In our reading, the syllabi already name evaluation, bias and privacy; what they do not single out is checking the everyday AI answers students meet. Our suggestion, for the school and not as a judgement of CBSE’s documents, is to make checking AI output against a source a habit.

What time, equipment and teachers does the AI subject need?

Time first: in each of Classes 9 and 10, code 417 runs to 210 hours, 160 of them in Part B, the subject-specific units.

For equipment, the Class 10 syllabus recommends quantities for a batch of 20 students with a human-machine ratio of 2:1, and recommends using devices from the school’s existing inventory. The Class 11 syllabus calls its list suggestive and says the institution should procure only basic tools, equipment and accessories.

Qualifications for teachers appointed on contract should be decided by the State or UT, the Class 9 syllabus says. Its suggestive minimums are a diploma, BCA or B.Sc. in computer science or IT, a graduate degree with a PGDCA, or a DOEACC A Level certificate, plus at least 1 year of work experience “in the same job role”.

CBSE’s 2026 notifications list includes “Capacity Building Programmes for teachers teaching Artificial Intelligence in Class XI & XII” (June), two more such programmes, offline (notified July and August), and a January notification titled “Inclusion of Trainings conducted by Department of Skill Education, CBSE in Mandatory Training Hours”.

UNESCO’s AI competency framework for teachers (2024) sets out 15 competencies across five dimensions, from a human-centred mindset to AI pedagogy and AI for professional learning, at three levels: acquire, deepen and create.

What can a school add beyond the syllabus?

Each strand below is our suggestion, tied to lines in CBSE’s syllabi and UNESCO’s documents. UNESCO’s AI competency framework for students (2024) sets out 12 competencies across four dimensions at three levels: understand, apply and create.

What does more building add?

Class 11’s capstone stages include understanding the users and designing the solution. UNESCO’s framework names AI system design, with “Create” as its third level. Our suggestion: more than one small build a year, each ending in a test on other students and a written list of what failed.

What does checking outputs add?

UNESCO’s framework emphasises critical judgement of AI solutions. A school can apply the syllabi’s confusion-matrix tally to a general-purpose AI chatbot’s answers in an ordinary subject, so that checking machine output against the textbook becomes routine outside the AI period too.

What do data and privacy rules add?

Class 9 asks students to learn measures that protect data privacy, and Class 12’s computer vision unit lists privacy among the challenges to analyse. UNESCO’s Guidance for generative AI in education and research (2023) says most countries’ lack of generative AI regulation leaves users’ data privacy unprotected. A school can go further and make privacy a rule in every AI lesson, beyond a topic in the syllabus.

What could three AI lessons look like in class?

These are our worked examples for a school to adapt, not CBSE activities, and they claim no outcome. Each one uses made-up or public data only.

Example A, Class 8, “Rules or examples?” (about two periods, no devices). Each group gets 20 made-up picture cards, such as leaves drawn with a few features. One group writes a sorting rule; the other learns only from 10 labelled cards. Both sort 10 new cards, tally right and wrong, and discuss where each method failed.

Example B, Class 9 or 10, “Check the machine’s answer” (two periods). Through the school’s account, the teacher puts 10 questions from a finished chapter to a general-purpose AI chatbot. Pairs check each answer against the textbook, mark it right, wrong or not checkable, and tally the marks in a grid set out like a confusion matrix, then rewrite one wrong answer with its textbook page.

Example C, Class 11 or 12, “Build it, then test it on people” (four to six periods over two weeks). Pairs use an AI tool to build something small, for example a quiz on one chapter that stores no real data. They try it on 5 classmates, log every error, and write a one-page note: what it does, what data it uses, where it failed and who could be misled.

Table 2. Our three worked examples, not CBSE material, for a school to adapt, using made-up or public data only; no outcome is claimed.

ExampleClassWhat it addsPeriodsWhat students hand inWhat the teacher checks
A. Rules or examples?8Data: how examples, rather than rules, can teach a machine; links to pattern recognition and data interpretation, and to Class 9’s rule-based and learning-based modelsAbout 2, no devicesEach group’s tally and one sentence on where its method failedBoth groups sorted the same new cards; no card shows a person
B. Check the machine’s answer9 or 10Checking outputs against a source; links to true and false positives, Class 10’s confusion matrix and UNESCO’s critical judgement2The marked answers, the tally grid and the rewritten answerEvery “wrong” carries a page reference; the prompts held no personal details
C. Build it, then test it on people11 or 12Building, then data and privacy; links to the capstone’s stage on understanding users, Class 12’s hands-on use of AI tools and UNESCO’s AI system design4 to 6, over two weeksThe working tool, the error log and the one-page noteThe tool stores no real data; every logged error appears in the note

Placement is our suggestion too: A inside a maths or science period in Class 8; B after Class 10’s unit on evaluating models, or with Class 9’s evaluation terms; C alongside the capstone.

What should the school check before students use an AI tool?

Age comes first. UNESCO’s press release of 7 September 2023 says its Guidance for generative AI in education and research sets an age limit of 13 for the use of AI tools in the classroom. That is UNESCO’s recommendation, not Indian law or a CBSE rule.

The checks below are our suggestions, not law:

  • No real names, photos, voices or contact details go into any prompt.
  • Students use tools only through accounts the school sets up, under its own policy.
  • The teacher reads each tool’s own terms for a minimum age before students use it.
  • Student work stays on the school’s own systems.
  • Parents hear about the lessons in advance, through a note like the one below.

The Press Information Bureau’s release on the Digital Personal Data Protection Rules (14 November 2025) says an organisation handling data must get consent that can be verified before it processes a child’s personal data, though it allows limited exemptions for essential purposes such as healthcare, education and real-time safety. It adds that the Rules provide an 18-month phased compliance timeline.

Whether that duty applies yet, or an exemption reaches a given tool or the school, is for counsel.

Our school guide quotes the consent sections of the Digital Personal Data Protection Act itself. This is not legal advice; ask the school’s own counsel.

What should a note to parents about AI lessons say?

A note before the first lesson can tell parents what the class will do, with which tools, and what never goes into one. The template is ours and promises no result.

Where One Young India fits

We run programmes in this area, so weigh this section with that in mind.

Schools can join the Learnacy Schools network, which One Young India runs and whose page says “A network, not an association”. It costs nothing: no membership fee, no dues and no franchise. A school applies by describing its present teaching for curiosity; One Young India replies with the charter and the seal, and a short conversation follows. Become a partner school.

We do not pay schools, principals, teachers or other staff for referring students.

One Young India’s Curious Computing is a self-paced course for students: 26 chapters in four parts, about 12 hours, including a part on what AI is and is not, and its limits. It never asks for a password or any real personal data, and names no company or app to praise or blame. The welcome and chapter 1 are free; Learnacy Essential opens every chapter. Start free.

Learnacy Essential pays for the tools, the courses and the feedback. It never buys acceptance.

One Young India’s Entrepreneurship in Action is for students in Grades 8 to 12 who want to build with AI and run their own ventures; every participant builds and ships a working product and a founder’s pitch. No coding needed: students use “vibe coding”, directing AI in plain language. It includes a year of Learnacy Essential. Dates and fees are on the programme page.

  • Our pages do not state: which classes Curious Computing is written for, whether a school can enrol a whole class, or whether a teacher can see students’ chapter checks. Ask us in writing before agreeing.
  • For the programme above, the page does not state which AI tools students build with, whose accounts they use, or what device a student needs at home; ask before paying.

Common questions

Is AI compulsory in CBSE schools?

For Classes 3 to 8, CBSE’s Secondary Curriculum 2026-27 says schools are directed to introduce computational thinking and AI, inside existing subject periods. For students in Class 9 in 2026-27, Acad-70/2026 says the CT and AI modules in Class 9 are not part of their curriculum. In Classes 9 to 12, Artificial Intelligence is a skill subject a student may choose.

What is CBSE’s AI subject code for Class 9 and Class 11?

Code 417 covers Classes 9 and 10, and code 843 covers Classes 11 and 12. Each carries 100 marks, 50 theory and 50 practical.

At what age can students use AI tools in class?

UNESCO’s 2023 guidance on generative AI sets an age limit of 13 for the use of AI tools in the classroom, its press release says: a recommendation, not Indian law. Beyond that, the school’s own policy and each tool’s own terms apply.

Does a school need its own AI policy?

UNESCO reported in 2023 that, in its global survey of over 450 schools and universities, less than 10% had institutional policies or formal guidance on generative AI. Our suggestion is a short written policy built from the checks above.

How can teachers prepare to teach AI?

UNESCO’s AI competency framework for teachers (2024) maps 15 competencies across five dimensions at three levels. CBSE’s 2026 notifications list carries capacity building programmes for teachers teaching Artificial Intelligence.

Sources

Read on 16 November 2026 (India time) for other publishers’ pages, and between 8 and 16 November 2026 for our own pages. CBSE’s syllabi and its circular are cited through the CBSE pages that list them. How this page is checked and corrected: the editorial desk.

  1. Central Board of Secondary Education (CBSE), CBSE Academic, Computational Thinking and Artificial Intelligence (CT&AI): the CT and AI curriculum for 2026-27 and the handbooks for Grades 3 to 8.
  2. CBSE Academic, Circulars: Acad-70/2026 (29 October 2026) and the 2026 notifications on capacity building and training hours.
  3. CBSE Academic, Curriculum for the Academic Year 2026-27: the Secondary Curriculum 2026-27 (Part 1) and the skill subjects.
  4. CBSE, Department of Skill Education, Skill Education: the 2026-2027 curricula for Artificial Intelligence, code 417 (Classes IX and X) and code 843 (Classes XI and XII).
  5. Press Information Bureau (PIB), Ministry of Education, release on CBSE’s curriculum on Computational Thinking and Artificial Intelligence, 1 April 2026.
  6. Press Information Bureau (PIB), release on the Digital Personal Data Protection Rules, 2025, 14 November 2025.
  7. UNESCO, AI competency framework for students, 2024.
  8. UNESCO, AI competency framework for teachers, 2024.
  9. UNESCO, Guidance for generative AI in education and research, 2023.
  10. UNESCO, press release, “UNESCO: Governments must quickly regulate Generative AI in schools”, 7 September 2023.
  11. One Young India (ours): the Learnacy Schools page, the Curious Computing page and preview and the programme page for students who build with AI.

Related

For schools

Bring a Model G20 summit to your school.

Model G20 lets a school host its own economic-diplomacy summit, built around the G20’s agenda and decided by consensus: One Young India supplies the committees, rules and training, and the students run the summit. It is an independent educational programme, not affiliated with the G20.

The Learnacy Schools network is free to join: no dues, no franchise.

Schools that have hosted a Model G20 are listed edition by edition. See schools that host