Google turns Gemini into a mentor for Indian tech labs

AI enters the curriculum through the back door

Google, in collaboration with the government agency NITI Aayog, is launching ATL Saathi in India. It is a Gemini-powered web application serving as an assistant for teachers in Atal Tinkering Labs (ATL) (technology classrooms focused on 3D printing, robotics, and the Internet of Things). The tool helps educators prepare lessons, generate learning materials, and answer specific student questions.

The application is currently rolling out to a hundred pilot schools. Beyond generating summaries and quizzes for twelve core modules, Gemini 3.5 Flash can prepare specific project ideas, wiring diagrams, and safety instructions tailored to what students want to build. It handles this in eight local languages.

Controlling the infrastructure means setting the standard

This is a strategic shift for Google. While most companies try to sell AI tools directly to individual schools, here Gemini is integrated into an existing nationwide network of labs accessible to over ten million students. This is not just a software sale; it is about establishing the default ecosystem for an entire upcoming generation of Indian engineers.

If teachers learn to rely on a Google AI assistant for lesson prep and solving technical roadblocks in robotics, Gemini will become the invisible backbone of technical education. For Google, this means massive data collection on how people learn technical subjects and the creation of a habit that students and teachers will carry forward.

An assistant, or a replacement for missing experts?

The reality of the Indian education system is that there are not enough qualified teachers for advanced robotics and programming. ATL Saathi pitches this as relief from administrative work, but it actually addresses a deeper problem: AI acts as a proxy mentor where a human with the necessary know-how is missing.

The risk lies in what happens when the model generates an incorrect hardware wiring diagram. A teacher who does not deeply master the subject will not spot the error. The model bears no responsibility, and reliance on it could lead to the mistake being replicated across the network.

Adoption beyond the pilot will be the deciding factor

The proof of whether this bet pays off will not be the hundred pilot schools, which likely have premium support. The key metric will be whether the application starts spreading organically to places where Google is not on standby.

If it turns out that teachers using ATL Saathi can actually deliver higher-quality technical lessons, we will see similar government integrations in other countries. This gives Google a massive advantage over OpenAI, which struggles much more to penetrate government education structures.

Lilith's verdict

The charity facade hides a strategic fight over who gets to force their cloud and models as the operating system for state education.