China’s move to make AI education compulsory in primary schools sets a global benchmark for ambition.
Credit: iStock Photo
Punjab and Odisha have taken remarkable strides in including artificial intelligence (AI) in school education. Punjab’s education department has become the first in India to roll out an AI curriculum across its government schools. At the same time, Odisha’s ‘AI Policy 2025’ envisions AI instruction in 35% of its schools by 2029 and 90% by 2036.
These are baby steps, though, when compared to China, which has announced a nationwide mandate to introduce AI education in all primary schools beginning in September.
These examples signify a crucial global shift: AI is now integrated into children’s earliest learning experiences, and is no longer confined to higher education or specialised training. But for India, a key question looms large: Is it truly ready to follow this path toward universal AI education?
Punjab’s initiative marks a bold experiment in public education reform. The state’s AI curriculum is designed as a three-year rollout prioritising hands-on, project-based learning. Students are encouraged to engage in coding exercises, AI hackathons, and science fairs where they apply machine learning principles to real-world challenges. The state’s digital infrastructure gives it a strong head start, with over 95.6% of 19,243 government schools equipped with functional computers, one of the highest rates in India.
Odisha’s trajectory blends long-term planning with an inclusive vision. The Odisha AI Policy 2025 outlines a decade-long roadmap to integrate AI learning into the state’s education system. Building on the ‘Odisha for AI’ and ‘AI for Youth’ initiatives, the policy promotes digital literacy through free online courses, AI labs, and specialised training modules for teachers.
The plan also taps into the India AI Mission’s ‘Future Skills’ programme to equip schools with infrastructure and curriculum content. Odisha’s approach highlights the importance of linguistic diversity, with AI lessons tailored in Odia and tribal languages, demonstrating a commitment to equity in technology.
Punjab and Odisha offer two distinct but complementary models: one grounded in strong infrastructure and immediate implementation, the other in long-term inclusion and adaptability. Yet, they are outliers in a country where most school systems remain unprepared for such technological leaps.
Chinese global benchmark
China’s move to make AI education compulsory in primary schools sets a global benchmark for ambition. Backed by substantial research, infrastructure, and teacher training investments, China aims to establish universal AI literacy by embedding computational thinking and data science into early education.
Its top-down governance model allows swift deployment, and its standardised education system ensures nationwide adoption. However, this model presumes a baseline of digital competence and resource parity, conditions starkly missing in India.
Despite policies like the National Education Policy (NEP) 2020, and the forthcoming plan to introduce AI in Indian schools from Class 3, starting in 2026-2027, India’s readiness gap remains vast. Data from the Department of School Education shows that for 2023-2024, only 57.2% of schools in India had computers, and 53.9% had Internet access.
In poorer states such as Bihar and West Bengal, those numbers drop below 25%. A 2021 Azim Premji Foundation report reveals that nearly 60% of children could not access online learning even during the pandemic. Only 26.8% of Indian youth in the academic age group between 6-14 possess basic Internet browsing skills based on a 2021 survey, conducted by policy think-tanks LIRNEasia and ICRIER, with Meghalaya and Tripura reporting figures below 10%. Such low digital literacy rates, especially in rural areas, are a significant roadblock since AI education presumes a baseline of digital competence across teachers and students.
While proponents of AI literacy envision democratised access to high-value skills and global competitiveness, making AI education mandatory could inadvertently widen the existing education gap between rural and urban India. Private and elite urban schools, already technologically advanced, will flourish under the new model, while resource-poor government schools risk falling further behind.
Weak regulatory frameworks
The financial burden of maintaining devices, connectivity, and teacher training will weigh heavily on underfunded education systems. Moreover, introducing advanced technology without robust policy safeguards invites significant risks. India’s regulatory frameworks for digital education and AI governance are inadequate to address student data privacy, algorithmic bias, and the ethical use of technology.
Without comprehensive data protection and precise accountability mechanisms, AI use in classrooms could lead to surveillance, profiling or discrimination, particularly of students from vulnerable backgrounds.
China’s assertiveness is rooted in its systemic uniformity and substantial financial resources, which are lacking in India’s diverse and fragmented educational system. India requires a more decentralised and context-specific approach with 250 million school students across 1.5 million institutions. Applying China’s model without making necessary adjustments could lead to institutional overreach and further marginalisation of those already excluded from the digital landscape.
Instead, Indian policymakers must cultivate a middle path that enables technological uplift while acknowledging structural inequalities. Odisha’s focus on localised language content and phased rollout shows how federal flexibility can co-exist with a futuristic vision. Punjab’s model demonstrates how targeted State-level infrastructure investment can effectively ground an AI curriculum. Scaling these examples nationally demands not speed, but careful sequencing.
A responsible national strategy should begin with three foundations. First, the Union government must prioritise equitable access to electricity, Internet connectivity, and hardware across all regions. Investments should extend beyond urban clusters into rural and tribal school networks. AI pedagogy requires teachers who are digitally literate, and ethically equipped to guide students through complex technologies.
Pilot projects under CBSE’s framework to train 10 million teachers nationwide are an essential start. Finally, clear national standards must address data protection, algorithmic fairness and accountability in educational AI systems. Without these, India risks creating a generation that learns through systems it cannot question or control.
India’s strength lies not in mimicking China’s rapid rollout, but in designing an AI education model rooted in fairness and adaptability. The challenge is to bridge aspiration with capacity — to ensure that digital innovation does not deepen existing inequities.
Initiatives like NITI Aayog’s Sath-E that encourage states to reorganise and merge schools with low enrolment, have spiked student dropouts in remote areas due to the closure of schools.
Instead, taking inspiration from cases, as in Japan, which ran a train for just one student to attend school, the primary focus should be on providing last-mile access to basic education for every child.
As India prepares to mainstream AI from Class 3 by 2026, it stands at a critical crossroads. It can choose to replicate others’ urgency or build the foundations of its own version of inclusive technological learning.
The potential is enormous, but so are the risks. Unless innovation is combined with inclusion by training teachers, closing and safeguarding student rights, the AI revolution in education could shift from a promise to a threat.
In this turning point lies India’s real test: not whether it can teach AI, but whether it can teach it justly.
Namesh Killemsetty is Associate Professor, Jindal School of Government and Public Policy, O P Jindal Global University, Sonipat, Haryana.
This article was first published under Creative Commons by 360info.
(Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH)