Sneha Sharma
Abstract
Generative AI has moved from novelty to default study habit in Indian classrooms, raising a question policy has not yet fully answered: what happens to students’ thinking when AI starts doing the thinking for them? This article looks at the emerging evidence on AI-driven “cognitive-offloading”, places it against India’s own foundational learning and employability data, examines the country’s current policy response, and outlines where gaps still lie and what closing them could look like.
Introduction
When one walks into any Indian classroom preparing students for board exams today, and you will likely find AI chatbots doing quiet, constant work such as drafting essays, solving math problems, and summarising chapters, which was not the same until 2022, when students were seen deeply involved in class lectures and books to learn new concepts. Now it is no longer a fringe habit. It is becoming the default way an entire generation studies. The question now India has to confront is whether AI belongs in education.
That decision has effectively already been made, but what happens to students’ thinking when
the tool that is supposed to help them learn starts doing the learning for them.
The Problem: Why This Is Different Global research has begun documenting a pattern researchers call “cognitive offloading”, which is the tendency to hand mental effort over to an external tool rather than exercise it ourselves. Emerging studies
link heavy AI reliance to weaker independent reasoning and problem-solving, particularly among young users. In the United States, national surveys of school-going teens now show a majority worries about AI’s effect on their own thinking, and that concern has been rising through 2025 rather than fading as students grow more used to the tools.
The pattern is consistent enough across contexts to take seriously,
even if and this caveat matters in the long term; a casual picture is still being built by researchers and shouldn’t be overstated as a settled fact.
India’s situation adds a layer the global conversation often misses: AI-driven cognitive offloading is arriving on top of foundational learning gaps that already exist. ASER 2024, India’s largest citizen-led learning assessment, found that only about 23% of class 3 students in government schools could read a class 2-level text, and roughly two-thirds of class 3 students could not perform simple arithmetic.
These are pre-AI numbers; they describe a system already working to build basic reasoning skills. Layering rapid, unguided AI use on top of that foundation raises real questions about whether AI will help close these gaps or make them harder to see, since a student can now easily make a polished essay or solve algebraic problems without ever building the skill an exam is meant to certify.
At the other end of the pipeline, the employability data suggests the stakes are not abstract. The India Skill Report, 2025, which was based on over 6.5 lakh candidates and input from more than 1,000 employers, found even as technical hiring accelerates, graduates are consistently falling short on communication skills, critical thinking and judgement. Separately, India’s Graduate Skill Index foundIndian graduates score over 46% employability in “learning agility”, lower than their overall soft skills core.
India’s response so far To its credit, India is ignoring AI in education. The Ministry of Education, working with NCERT and CBSE, is rolling out an Artificial Intelligence and Computational Thinking Curriculum starting from class 3 in the 2026-27 academic year, under the framework of NEP 2020 and the National Curriculum Framework for School Education 2023. Teacher training is being built into this through NISHTHA, the government’s teacher development programme, and the stated goal is to make AI literacy as foundational
as reading and numeracy.
This is a meaningful step, but it is also, so far, a curriculum about AI, not a curriculum that protects thinking from overreliance on AI. Publicly available material on the rollout focuses on building AI skills and computational thinking as a subject area.
What’s less visible is any explicit safeguard against cognitive offloading in how AI is used across the other subjects or any direct link between this curriculum and specific skills, like critical thinking, judgement, and independent problem solving that national skilling data says employers are struggling to find. NEP 2020 itself calls for critical thinking as a core educational goal, but the operational bridge between “teach AI” and “protect independent reasoning while using AI” doesn’t yet appear to be built out in policy.
Way forward
Nothing here suggests that AI use in schools should be opposed, as it has already been accepted, and personalised learning and accessibility offer real advantages. However, what is needed is to take a more serious look at the way in which AI can be employed and consider it not only from the standpoint of whether it is being included in curricula. There are three lines to pursue. Firstly, pedagogical guidelines in which there will be clear demarcation between AI as scaffolding and AI as substitution in order to make sure that both teachers and pupils have the same understanding of acceptable usage.
Secondly, assessment reform in which tasks and exams will be structured in a way that will not allow full outsourcing to the chatbot and will retain the productive struggle necessary for the development of reasoning. And most importantly, direct connection of school-level AI curriculum to employability skills emphasised in the India Skills Report.
India, however, remains at an early and flexible stage of implementing such a policy: The AI and Computational Thinking course continues its development until 2026–2027; the teacher’s training modules have yet to be finalised, and the technical implementation of the connection between the AI literacy and critical thinking protection remains open for adjustment.
It is, in a way, good news, as it implies that the gap described above is still adjustable rather than something that needs to be adapted once implemented. Moreover, the current generation that uses AI to do its homework is the same generation which will need to demonstrate the very skills whose development may be hindered by the use of AI. It cannot be solved by the tech companies or the examination boards alone; it is a governance issue, and India has only a limited time to solve it before its AI in schools policy becomes unchangeable.
Reference
ASER Centre / Pratham Education Foundation. (2025). Annual Status of Education Report (ASER) 2024.Pratham https://asercentre.org/
Gerlich, M. (2025). AI tools in society: Impacts on cognitive offloading and the future of critical thinking.Societies, 15(1), 6. https://doi.org/10.3390/soc15010006
Mercer | Mettl. (2025). India’s Graduate Skill Index 2025. Mercer.
https://www.mercer.com/insights/talent-and-transformation/talent-assessment/indias-graduate-skill-index-2025/Ministry of Education, Government of India. (2020). National Education Policy 2020.https://www.education.gov.in/nep/about-nep
Ministry of Education, Government of India. (2023). National Curriculum Framework for SchoolEducation. NCERT. https://ncert.nic.in/pdf/NCF_for_School_Education_2023.pdf
Ministry of Education, Government of India, Department of School Education and Literacy. (2025).
Artificial Intelligence and Computational Thinking curriculum rollout from Class 3, academic year2026–27 [Press release/curriculum framework]. https://cbseacademic.nic.in/
Miao, F., & Holmes, W. (2023). Guidance for generative AI in education and research. UNESCO.https://doi.org/10.54675/EWZM953
RAND Corporation. (2026). More students use AI for homework, and more believe it harms critical thinking: Selected findings from the American Youth Panel.
https://www.rand.org/pubs/research_reports/RRA4742-1.html
Wheebox, Confederation of Indian Industry (CII), All India Council for Technical Education (AICTE), & Association of Indian Universities (AIU). (2025). India Skills Report 2025. Wheebox.https://www.wheebox.com/india-skills-report.html
World Economic Forum. (2025). The future of jobs report 2025.
https://www.weforum.org/publications/the-future-of-jobs-report-2025/
About the Contributor
Sneha Sharma is a master’s in economics student at Gokhale Institute of Politics and Economics with a strong interest in the development sector relating to education, labour force, gender and public policy. She is currently a research and editorial intern at IMPRI, Impact and Policy Research Institute, where she contributes to policy update articles and research focused on strategic affairs and policy initiatives.
Disclaimer
All views expressed in the article belong solely to the author and not necessarily to the organisation.
Read more at IMPRI:
Rural Women’s Property Revolution : NHFS-6 on Women Property Ownership
Thirteen years after Delhi gangrape, women’s safety is still a huge issue
Acknowledgement: This article was posted by Prisha Sachdeva, a Research and Editorial Intern at IMPRI.
