Picture an Academic Director at a CBSE school in Nashik. This month she has done what the state asked. Her teachers are enrolled on AI Saathi, the four-hour AI course Maharashtra has made mandatory, and the deadline has just moved again. Then she opens the timetable for Classes 3 to 8 and counts the hours that need a name beside them. In a school with two sections per class, she finds 900.
That gap is what this post is about. AI teacher training in Maharashtra schools is moving fast, and it is genuinely useful. But training teachers to use AI and staffing a lab timetable are two different jobs.
What Is AI Saathi, and Who Has to Complete It?
AI Saathi for Educators is a self-paced course on DIKSHA, built by SCERT Maharashtra with Kotak Education Foundation and written entirely in Marathi. Education Minister Dadaji Bhuse launched it on 14 July 2026. According to The CSR Universe, it is the first mandatory virtual course under the state's online training policy. It counts for four hours of the state's 20-hour online training framework and is compulsory for nearly 7.5 lakh teachers and education officers from Classes 1 to 12. It has three levels, Basic, Intermediate and Advanced.
The deadline has moved twice. SCERT first asked teachers to finish by 31 August, then by 30 September. The latest circular, reported on 6 October, gives teachers until 31 October to register and 30 November to complete the course.
Who is covered? State notices describe the course as meant for all teachers of Classes 1 to 12, and the revised circular is addressed to divisional deputy directors, extension officers, district training institutes and municipal education officers. That is the government school system. Whether the mandate binds a CBSE or private unaided school is a question for your district education office, and we would ask it in writing. Either way, the date is a useful prompt to plan your own training calendar. Our post on what teachers need after NISHTHA covers the national picture. If your school follows the Maharashtra state board and is moving to the CBSE pattern, read our plan for that shift alongside this post.
What Is the Google AI Educator Series Programme for Maharashtra Teachers?
In June, the School Education Department signed a Letter of Intent with Google for Education to train more than four lakh teachers over 18 months, as Akashvani reported. Master trainers go first and then train teachers through the Google AI Educator Series, free of cost to the state. Minister of State Pankaj Bhoyar framed the programme for teachers in government and aided schools, with content in Marathi, Hindi and English.
The rollout runs 18 months and starts with master trainers. A school waiting for it to reach every teacher is waiting a long time. And because the announcement frames the programme for government and aided schools, a private school that wants its teachers on the Google series should ask its district office how, or whether, they can join.
What Do These Programmes Teach, and What Do They Leave Out?
Both programmes do one job well: they make teachers comfortable with AI tools. A Marathi news report on the Basic level lists AI fundamentals, lesson planning, making educational videos and using tools such as ChatGPT, Gemini or Canva AI. Other summaries add study material and assessment. That is teacher productivity plus AI literacy, and it is worth having. A teacher who plans lessons with AI gets hours back.
But look at what the descriptions we could find do not mention: robotics kits, electronics, sensors, drones, EV systems, AR/VR headsets. Nothing says a teacher will walk out able to run a full period of hands-on work with a batch of 30-plus students and a box of components. That is a different skill. It means managing hardware, fixing a motor that will not spin and keeping a batch safe, and a four-hour video course cannot teach it.
Picture a Class 7 session on sensors. There are eight kits for 32 students. One breadboard has a loose jumper wire, another has an LED the wrong way round, and a third group has connected the power backwards. The session depends on a person who can spot all three in two minutes and keep the rest of the batch moving. A teacher can learn that, but not from a video course.
AI literacy tells a teacher what AI is. A lab timetable needs someone who can run the session.
How Many Hours Does a School Actually Have to Teach?
Count it yourself. CBSE's Computational Thinking and AI curriculum for Classes 3 to 8 began in 2026-27. According to summaries of the framework, it carries about 50 hours a year in Classes 3 to 5 and 100 hours a year in Classes 6 to 8. We unpack the structure in our posts on the CBSE resource books and school setup for CT&AI. Now take a school with two sections in every class:
- Classes 3 to 5: 3 classes, 2 sections, 50 hours each, so 300 hours
- Classes 6 to 8: 3 classes, 2 sections, 100 hours each, so 600 hours
- Total: 900 hours of CT&AI teaching a year
Add the Composite Skill Lab requirement. CBSE's guidelines expect 110 hours a year of skill education in Classes 6 to 10, which is another 1,100 hours for the same school (5 classes, 2 sections, 110 hours each). We cover that side in our post on the CSL teacher training gap.
To be fair, not every one of those hours needs a lab. The early CT&AI hours in Classes 3 to 5 are largely unplugged and activity-based, and a teacher who has finished AI Saathi can run many of them. That is where the course helps most. The hands-on hours in Classes 6 to 8 and the skill lab are where delivery capacity runs out.
The hours are also growing. CBSE has said NCERT will provide CT&AI modules for Classes 9 to 12, and our timeline explainer on the Class 9 and 10 mandate sets out the dates schools are working to.
So who teaches those hours? Four hours of training does not answer that. Nor was it ever meant to.
What Should an Academic Director Do Before 30 November?
Start with the paperwork, then the timetable.
- Confirm coverage: ask your district education office or board contact, in writing, whether your teachers must complete AI Saathi and whether the four hours count toward your CPD record. CBSE has also named CT&AI its 2026-27 teacher training theme, so its workshops are another source of CPD hours.
- Enrol your teachers anyway, if they have not started. Comfortable teachers are a real asset for the unplugged hours.
- Write your own hours map: every class, every section, the CT&AI and skill education hours, and a teacher's name or the word "vacant" beside each line.
- Decide who covers the vacant hours: train more of your own teachers, hire specialist trainers, or partner with a lab provider that places an engineer on campus.
The 2026-27 session is already underway, so every week that a vacant line has no name beside it is a week of curriculum hours at risk. Each route has a cost and a timeline. We compare them in our post on the STEM lab teacher training problem and describe the third route in what a good on-campus lab trainer does. In our model, an on-campus engineer conducts the sessions, and a Faculty Development Program with NEP 2020 CPD hours builds your own teachers' confidence alongside. You can see how we approach the specialised teacher shortage in full.
Frequently Asked Questions About AI Teacher Training in Maharashtra
Is AI Saathi mandatory for private school teachers?
State notices describe it as mandatory for all Classes 1 to 12 teachers in the state, and the revised circular went to state and municipal education officers. Whether it binds private unaided or CBSE schools is not clear from what we have seen. Ask your district education office in writing.
What is the AI Saathi deadline?
As of mid-October 2026, teachers can register until 31 October and complete the course by 30 November 2026, after two extensions. Check DIKSHA and SCERT circulars, because the dates have moved before.
How many CPD hours is AI Saathi worth?
It is worth four hours within the state's 20-hour mandatory online training framework.
Does AI Saathi teach teachers to run a robotics or AI lab?
No. The course focuses on AI literacy, classroom integration and using AI tools for lesson planning, study material and assessment. The descriptions we found do not cover hands-on kits such as robotics, electronics or drones.
Who teaches AI lab classes if teachers are not trained for it?
Schools have three realistic routes: train their own teachers, hire specialist trainers, or work with a lab partner that places an engineer on campus. A school can also combine two of them.
Start With the Hours, Not the Course
Go back to the Academic Director in Nashik. Her teachers will finish AI Saathi before 30 November, and that is a good outcome. It still leaves 900 hours with no name beside them. The courses build comfort with AI. The timetable needs people who can run sessions.
So write the hours map this week and let it tell you what kind of help you need. When you want to see how a partnership covers the vacant hours, start with the full curriculum or see how the lab works.