It is July. The 2026-27 academic session is well underway. And in a Class 5 classroom in Nashik, a Marathi and EVS teacher named Mrs. Kulkarni is standing in front of 34 students, trying to run a pattern-recognition activity from the new Computational Thinking curriculum.
She completed her NISHTHA module on DIKSHA back in May. It took about three hours, spread across two evenings. She watched the videos, read the material, passed the quiz. Nobody told her she would also need to improvise when a student asked why the "pattern" in the worksheet had two possible right answers.
This is happening in thousands of classrooms right now. Not because anyone did anything wrong. Because a self-paced training module and a live classroom are two very different things, and India just asked 27,000-plus CBSE schools to bridge that gap in a single academic year.
What NISHTHA Training For AI and CT Actually Covers
NISHTHA, the National Initiative for School Heads' and Teachers' Holistic Advancement, is the government's primary vehicle for delivering teacher training on the new Computational Thinking and Artificial Intelligence curriculum. For the 2026-27 rollout, CBSE has designated "Computational Thinking and Understanding AI" as the official training theme, and it flows through three channels: self-paced modules on the DIKSHA platform, in-person District Level Deliberations (DLDs) carrying 3 to 6 CPD hours, and regional workshops at CBSE Centres of Excellence for a ₹700 registration fee.
We covered the compliance side of this rollout in detail in our guide to the CBSE CT and AI curriculum for Classes 3 to 8, including what schools need to set up before the session starts. This post picks up where that one leaves off. Setup is done. The session has started. So what does a teacher actually do on a Tuesday morning with this training behind her?
Here is the honest answer: NISHTHA gives teachers exposure. It does not give them practice.
Exposure Is Not Practice, and the Difference Shows Up Fast
A three-hour module can explain what computational thinking is. It can walk through sample activities, show a demo lesson plan, and quiz a teacher on definitions. What it cannot do is put 34 real Class 5 students in front of her, each moving at a different pace, each asking a slightly different version of the same question.
Think about what CBSE has actually structured here. For Classes 3 to 5, computational thinking is not a separate subject. It is folded into existing lessons, Maths, EVS, language, taught by the same subject teachers who were already running a full timetable. These are not computer science teachers. Most of them have never written a line of code and were never expected to. Now they are running "unplugged" logic and sequencing activities on top of everything else they teach.
For Classes 6 to 8, the model shifts again. Teachers from different disciplines are expected to collaborate on delivering CT and early AI concepts together. That sounds sensible on a policy document. In practice, it means cross-subject coordination that most schools have no existing structure to support.
Only by Classes 9 to 12 does the responsibility land where people assume it always does, with computer science teachers running dedicated AI instruction. For everyone below that, the training gap is structural, not personal. It is not that any individual teacher is unprepared. It is that the system asked generalist teachers to deliver a specialist subject with three hours of video and a quiz.
A national survey by CENTA found that around 70% of Indian teachers already use AI tools for lesson planning and classroom activities. That number tells you something important: teachers are not resistant to AI. They are comfortable using it as a tool. What NISHTHA does not solve is the much harder problem of teaching AI concepts to eight-year-olds in a way that holds their attention, corrects their misconceptions in real time, and builds toward something by December, not just survives the first two weeks.
The Infrastructure Question Nobody Mentions in the Training Videos
There is a second layer to this that rarely comes up in the training modules themselves. Roughly half of Indian schools still deal with unreliable internet, inconsistent electricity, or no dedicated computer lab access. NISHTHA's own design accounts for this to an extent. Classes 3 to 5 are meant to run on "unplugged" activities, printed worksheets, puzzles, physical pattern cards, so infrastructure gaps do not block the youngest learners.
But that is a workaround, not a solution. A teacher running an unplugged activity still needs to understand the underlying computational thinking concept well enough to adapt it on the fly when the printed worksheet does not quite land with her class. A module watched once in May does not build that kind of adaptive confidence. Confidence like that comes from repetition, from having someone in the room the first few times who can answer the question the teacher cannot.
We wrote earlier this year about how most STEM lab vendors treat teacher training as a single orientation workshop and then disappear. NISHTHA is not a vendor, so the comparison is not exact. But the structural failure mode is similar. A short training event, however well designed, cannot substitute for ongoing support once the actual teaching starts.
What "Supplemented" Actually Looks Like
So what should a school do differently, given that NISHTHA is already the government's chosen path and is not going away?
The honest answer is that schools need to stop treating NISHTHA completion as the finish line. It is the starting point. What comes after matters more.
A few things separate schools where this curriculum is actually landing from schools where it is quietly falling apart by October:
- A named point of contact for the teacher when she gets stuck, not a WhatsApp group that goes quiet after the first month.
- Session observation and light coaching, someone sitting in on early sessions and giving specific feedback, not generic encouragement.
- A buffer between subject teachers and the technical depth of AI content for the Classes 6 to 8 band, where cross-subject coordination is hardest.
- Continuity if the trained teacher is absent or transfers, which happens more often than school planning accounts for.
This is exactly the gap our on-campus engineer model was built to close. Our engineers do not replace what a subject teacher brings to a classroom. They sit alongside her, run the technical parts of the session directly when needed, and let her build real competence over the year instead of being thrown into the deep end after a three-hour module. We describe what a genuinely useful on-campus lab trainer does in practice, and the short version is: presence, every week, not a one-time visit.
Our Faculty Development Programme runs in parallel with student sessions and includes NEP 2020 CPD hours, structured around the reality of a teacher's existing full timetable, not a hotel conference schedule. You can see how the full curriculum maps across Classes 1 to 12, and how it is designed to support, not replace, the teachers already in the building.
Schools evaluating this for the first time, particularly CBSE schools working through the new compliance requirements, often assume the hard part is the curriculum content itself. It usually is not. The hard part is sustaining delivery quality from July through March, across a full academic year, with a teacher who has a hundred other responsibilities.
A Realistic Picture of the Rest of This Year
Here is what is likely to happen in most schools without additional support. The first month or two will go reasonably well, because the material is novel and the training is fresh. By Diwali, some sessions will start slipping, either simplified quietly or skipped when the timetable gets tight. By the end of the year, a subset of classrooms will have delivered a genuinely engaging computational thinking programme, and a larger subset will have delivered a watered-down version that technically checks the compliance box.
That is not a criticism of the teachers. Mrs. Kulkarni in Nashik did not choose to become a computational thinking specialist. She was asked to add it to an already full plate, given three hours of preparation, and told to make it work. Most teachers in her position are trying genuinely hard. The gap is not effort. It is support.
Schools that recognise this early, in July and August rather than in February when the gaps are already visible, have more room to fix it. That might mean structured internal mentoring between teachers. It might mean bringing in outside support specifically for the technical depth teachers were not trained for. Either way, the fix has to happen at the school level, because NISHTHA was never designed to be the whole answer. It was designed to be the starting point.
Frequently Asked Questions
Is NISHTHA training enough to prepare teachers for the CBSE AI and CT curriculum?
NISHTHA provides useful foundational exposure through DIKSHA modules and District Level Deliberations, but it is orientation-level training, not ongoing classroom support. Most teachers, especially subject teachers handling Classes 3 to 5 who were never trained in AI or coding, need continued mentoring and support once actual sessions begin.
Who is responsible for teaching AI and Computational Thinking in Classes 3 to 8?
For Classes 3 to 5, Maths, EVS, and language subject teachers integrate computational thinking into existing lessons. For Classes 6 to 8, teachers from different disciplines are expected to collaborate. Only from Class 9 onwards do dedicated computer science teachers lead AI-focused instruction directly.
What is the difference between DIKSHA self-paced modules and District Level Deliberations?
DIKSHA modules are online, self-paced training a teacher completes independently, typically a few hours of video and quiz content. District Level Deliberations are in-person sessions carrying 3 to 6 CPD hours, offering more direct interaction but still limited to a single event rather than ongoing support.
What can schools do beyond NISHTHA to support teachers with the new curriculum?
Schools can assign a named point of contact for questions, observe early sessions and provide specific coaching, build in cross-subject coordination support for the Classes 6 to 8 band, and plan for continuity if a trained teacher transfers or is absent. Some schools bring in a managed lab partner to provide this support directly, rather than building it internally from scratch.
Does NEP 2020 require ongoing teacher training, or is a one-time module sufficient?
NEP 2020 emphasises continuous professional development, not a single training event. While it does not prescribe an exact number of hours for the AI and CT curriculum specifically, the policy's broader intent supports sustained teacher development rather than a one-time orientation.
Mrs. Kulkarni will figure out her pattern-recognition activity this week. She always does. But the question worth asking, if you are a principal or academic director reading this in July, is whether you want her figuring it out alone, session after session, for the rest of the year. Or whether she has someone in the room with her when it gets hard.
If you are thinking through what that support should look like for your school, we are ready to talk.