Cat or dog?
An image classifier trained on photos students gather. Their dataset, their mistakes.
Skills · AI
AI, taught honestly.
Students build models, find where they break, and explain why. Classes 3 to 12, on computers you already have.
- Classes
- 3–12
- Form of work
- Any of the three
- CBSE circular
- Skill-81/2025
- Runs on
- The computer lab, in the browser
The pattern
What we've noticed
Most school AI teaching stops at the demo. We start with data students gather and end with what the model gets wrong.
The work
What students actually build
Our studio's exemplars, not student work.
Rain before sports day
A prediction model on public weather data. Misses presented with hits.
The bias hunt
Audit a ready-made classifier: find where it fails, and why.
Classes 3 to 12
Class by class.
Classes 3 to 5
unplugged and visual. CBSE's "Computational Thinking & AI" curriculum covers classes 3 to 8 from 2026-27.
Classes 6 to 8
three projects a year, one per form of work (Circular Skill-81/2025, 28.10.2025). A Composite Skill Lab "may utilize" it.
Classes 9 to 10
AI (subject code 417), an established skill subject. Elective, not compulsory.
Classes 11 to 12
bigger datasets, longer projects, portfolio depth.
Classes 3 to 12
Your computer-science teacher can run this. No machine-learning engineer required.
Training covers dataset sessions, teaching under real uncertainty, assessing process over product. By August to December teachers lead and we coach.
Equipment
Equipment is bought. Capability is built.
Starts on the computer lab you have: browser tools, no GPUs, no accounts for minors. CBSE's CSL equipment list is explicitly "suggestive", adoptable per readiness (Circular Skill-13/2026, 20.03.2026). Pedagogy first.
The skills map
Related skills.
Questions
Honest answers.
Is AI compulsory in classes 9 to 10?
No. Subject code 417 is an established skill subject, offered as an elective. The mandatory part is Kaushal Bodh, classes 6 to 8, session 2025-26 (Circular Skill-81/2025).
What about student data and privacy?
They need none: photos of pets and objects, weather records, anonymized lists. No accounts created for minors.
Can AI projects count toward Kaushal Bodh?
Yes. One project per form of work: life forms; materials and machines; human services. Assessment weights viva, activity book, portfolio, teacher observation.
Book a workshop
Half a day at your school. Here is what happens.
How a school and Hankernest find out whether AI fits before anyone commits to a year. Inside your timetable, with your teachers and your students, on what you already own.
in the demo period
Your class builds
A cat-or-dog classifier trained in the browser on photos the class brings, then shown the photo that fools it.
And learns. Their dataset, their mistakes.
- First period
A demo period with a class
One brief from this page, run as a lesson with one of your classes, on what the school already owns.
- Second period
Hands-on for your teachers
The same kit in their hands. Where it fails, what students ask, how the period is paced.
- Over tea
The timetable conversation
Where the periods come from, which classes, and which forms of work the projects would cover.
- Within the week
A written fit note
What we would run, what you already have, what you would buy directly, and a cost band. No obligation.
You pick the class and the period. We bring what the session needs.
Or just talk to us. info@hankernest.com, or +91-7000403607 on call and WhatsApp during office hours.