Somewhere in a staff room in Pune this month, a Maths teacher opened a PDF called CTAI_Pri6SH_2026-27.pdf and had no idea what to do with it. That is not a criticism. CBSE uploaded the Computational Thinking and Artificial Intelligence resource books for Classes 3 to 8 to cbseacademic.nic.in with very little fanfare, and most of what has been written about them since stops at "CBSE has released new AI resource books." Nobody has actually opened them up and explained what a Class 6 student is meant to do with Chapter 3.
We did. Here is what these books actually contain, how they are structured, and where the real planning challenge sits for schools.
What CBSE Published, and When
The books were notified through Circular No. Acad-18/2026, dated 9 April 2026, addressed to every CBSE-affiliated institution. This followed an earlier circular, Acad-15/2026, which had announced the curriculum itself on 1 April. The gap between the two tells you something: CBSE announced the mandate first and published the actual teaching material a week later. Schools that started planning off the announcement alone were working blind for those eight days.
The books are titled Computational Thinking and Artificial Intelligence, one per class from 3 to 8, and the first edition is dated March 2026. CBSE is explicit about what these are not. They are not textbooks. They are resource books, meant to sit alongside the existing Mathematics syllabus and be used by the Maths teacher, not a separate AI specialist. That single design choice explains almost everything else about how the content is built.
Part 1: A Companion to the Maths Textbook, Not a New Subject
Every book has two parts, and Part 1 is Computational Thinking. It follows the exact chapter sequence of the NCERT Maths textbook for that class. For Class 6, that means chapters like Patterns in Mathematics, Lines and Angles, Number Play, Data Handling and Presentation, Prime Time, Perimeter and Area, Fractions, Playing with Constructions, Symmetry, and The Other Side of Zero.
Inside each chapter, the questions are pure logic and reasoning puzzles. Grid problems, pattern sequences, symmetry riddles, the kind of thing you would find in a good reasoning olympiad book. There is no screen involved and no code. A Class 6 student might be asked how many diamonds appear in a sequence of hexagon-and-diamond patterns, or which square in a grid must be shaded to make it symmetric. The idea is that as the Maths teacher covers a chapter in the regular syllabus, she pulls out the matching CT chapter and uses it to deepen the same concept through a different kind of thinking.
This is a smart design on paper. It costs nothing extra in timetable terms because it rides on an existing subject. But it also means the quality of CT delivery depends entirely on how comfortable that Maths teacher already is guiding open-ended thinking, since CBSE explicitly tells teachers to prompt and question rather than hand out answers.
Part 2: Four AI Chapters, and They Get Denser Fast
Part 2 is where Artificial Intelligence lives, and for Class 6 it runs exactly four chapters:
- Introduction to AI and Everyday Examples, covering the meaning of AI, AI versus automation, human versus machine intelligence, and a first, simplified look at supervised, unsupervised, and reinforcement learning
- Basic Data Concepts, covering types of data (numerical, text, image, video, sound), how to collect it, and how to represent it using tables, bar charts, and pictograms
- Simple Pattern Recognition and Decision Making, on spotting patterns in everyday data and turning observations into conclusions
- Ethics and Digital Responsibility, covering plagiarism, hacking, phishing, spamming, privacy, software piracy, and digital footprints, built around a short case study of a student named Arjun who gets his files wiped after clicking a fake "you've won a game" email
That last chapter is worth pausing on. It is not a token paragraph about "being safe online." It walks through seven distinct categories of digital risk with concrete prevention steps for each, and it does so through story rather than definition. CBSE clearly wants ethics taught the way a good English teacher teaches a moral lesson, not the way a computer science teacher lists rules.
As classes progress from 6 to 8, this same four-chapter skeleton stays but the depth increases sharply. What is a definitional exercise in Class 6 becomes an applied one by Class 8, closer to actual data literacy work. Class 3 to 5 content is lighter still and largely activity-based, since NCF-SE assigns it only 50 hours a year against 100 for the middle stage.
What CBSE Deliberately Left Out
Here is the part most coverage of this circular skips entirely. CBSE is upfront that this curriculum is platform-agnostic by design. No coding platform, no app, no vendor tool is prescribed or required. The 100 hours for Classes 6 to 8 split as 40 hours Advanced CT, 20 hours Introductory AI, and 40 hours on two interdisciplinary projects, and none of that time assumes a computer lab, a robotics kit, or even a laptop per student.
That is a deliberate choice, and a reasonable one for a national mandate that has to work in a government school in rural Bihar and a well-funded private school in South Mumbai on the same day. But it leaves a real gap for schools with actual technology infrastructure sitting unused. A student who has just spent a chapter classifying supervised versus unsupervised learning on paper has learned the concept. She has not trained a model, written a line of Python, or seen what a dataset looks like outside a textbook diagram. CBSE's own resource books admit this by staying paper-based. They are not pretending otherwise.
There is also no exam pressure built in anywhere. Assessment is meant to be continuous and competency-based: project presentations, reflective journals, observation-based rubrics, not a written test with marks that go on a report card. For a teacher used to the security of a marking scheme, this is a genuinely bigger shift than the AI content itself.
How This Connects to What Happens in Class 9
This curriculum for Classes 3 to 8 does not exist in isolation. The same set of April 2026 circulars discontinued the standalone AI skill subject, Code 417, for Class 9 from the 2026-27 session onward, while Class 10 students already enrolled continue on the earlier scheme for one more year. We covered what that means for schools running Class 9 electives in our piece on the Class 9 skill-subject mandate. The eventual plan is for a unified, compulsory CT&AI track to extend into Classes 9 and 10 from 2027-28, with NCERT developing internal assessment modules for the senior years in the meantime.
So a school building its technology roadmap right now needs to hold two timelines in its head at once: the Classes 3 to 8 resource books that are live today, and a Classes 9-10 framework that is still being finalised. We walked through the operational side of getting the Classes 3-8 rollout timetabled correctly in our earlier post on setting up for the CT&AI curriculum, which is worth reading alongside this one if your academic team hasn't assigned periods yet.
What This Actually Asks of a School
Read plainly, CBSE's resource books ask for three things: a Maths teacher willing to guide rather than instruct, roughly two and a half hours a week found somewhere in an already full timetable, and a school comfortable assessing "AI literacy" without a mark sheet to point to. None of that requires new hardware. But none of it builds the hands-on capability that a parent means when they ask whether their child is "learning AI."
That gap between paper-based literacy and applied skill is exactly where our own 10-year AI and Robotics curriculum starts, and it is why we built it to run through an on-campus engineer rather than a workbook. A school in Nagpur running CBSE's resource books correctly is doing the compliant thing. A school pairing that with an actual project cycle, where the classification concept from Chapter 1 turns into a real image-recognition demo a student can show a parent, is doing the differentiated thing.
Frequently Asked Questions
Are the CBSE CT&AI resource books mandatory for all schools?
Yes, for every CBSE-affiliated school, from the 2026-27 academic session, for Classes 3 to 8. They were notified through Circular Acad-18/2026 and Circular Acad-15/2026.
Do the resource books replace the Maths textbook?
No. Part 1 of each book is designed as a companion to the existing NCERT Maths textbook, following its chapter sequence. It supplements the syllabus rather than replacing it.
Is coding part of this curriculum?
Not directly. The curriculum focuses on Computational Thinking and conceptual AI literacy. It is explicitly platform-agnostic, so it does not prescribe any coding language, app, or tool.
How many hours does the curriculum require?
Classes 3 to 5 require 50 hours a year. Classes 6 to 8 require 100 hours a year, split as 40 hours Computational Thinking, 20 hours introductory AI, and 40 hours across two interdisciplinary projects.
What happens to the AI subject in Class 9?
The earlier standalone AI skill subject, Code 417, is discontinued for Class 9 from the 2026-27 session. Class 10 students already on the older scheme continue it for one more year, ahead of a compulsory CT&AI track expected from 2027-28.
Will this curriculum be graded with marks?
No. CBSE has designed the assessment to be continuous and competency-based, using project presentations, reflective journals, and rubric-based observation rather than a conventional written exam.
Where This Leaves You
CBSE has done the hard part of writing a curriculum that works in any school in the country regardless of budget. That is precisely why it stops short of anything hands-on. The resource books are the floor, not the ceiling, and reading them chapter by chapter is the fastest way to see exactly where your school's own lab work should pick up.
If your academic team is still deciding what sits on top of this compliance layer, we would like to be part of that conversation.