By the time this article hits the stands, India AI Impact Summit would have come to a successful close with thousands returning with millions of ideas to disrupt the lifestyle of billions using silicon agents working in tandem with synaptic naturals. The convergence of artificial and natural has come at a time when the entire world is navigating the waters of AI wave that follows its predecessors—internet, e-commerce, social media. To carelessly stay away claiming tradition is best or to mindlessly merge and embrace modern algorithms is education’s dilemma. The abuse of AI in applications other than education is reversible while the same is irreversible in education especially during its formative stages. It is at this crossroad of policy conundrum that education seeks a good mix of algorithmic assistance and cerebral persistence to achieve Prime Minister Modi’s target of India being among top three global AI super powers by 2027.
Global data suggests that the use of AI in education is incredibly widespread at an adoption rate higher than in most other industries. With an 86 per cent adoption rate, the global AI in education market of USD 7.57 billion in 2025 is expected to touch USD 112 billion by 2034. While 88 per cent students using agentic AI for academic work, 60 per cent of teachers using AI for organised pedagogy with students using predominantly for information retrieval and homework/assignment assistance and faculty for lesson planning and assessment. In addition to the two, institutional administrators use AI for analytics to drive admissions, administration and accountability. The comity of students, faculty and institutional administrators present an interesting mix of opportunities and challenges, especially during this admission season where AI-based courses in colleges seem to mercilessly operate in the ‘highest bidder is the ultimate winner’ mode. I reserve a detailed article on this later with AI education becoming a booming commercial proposition and get back to my concern on AI in education, which is a civilisational juxtaposition.
The future of education is not in the outcome of a binary war between artificial intelligence and traditional pedagogy but in the calibrated adoption of both. While silicon-powered algorithms increasingly arbitrate attention, epistemic authority still relies on synaptic triggered realisms that are human-centric and not driven by agentic AI. The ground reality that formative education cannot be reduced to pattern recognition instruments or probabilistic optimisation models but an apprenticeship in ambiguity for limitless exploration is the natural guardrail for AI adoption in education. The need for architecting an intelligent synthesis of an AI-powered system to nurture human cognition is tellingly visible and cannot be ignored through blindfolded policies.
The AI scaffolding in education undoubtedly is improving access to education which was once the privilege of a few or addressing the challenging contours of learners with a huge baggage to be unburdened. However, digital temperance for students is non-negotiable as students should wrestle unaided and not forfeit their opportunity to build their meta-cognitive resilience. Resorting to a quick-fix solution deploying AI derails the youthful and formative stages of cerebral salience. Equally important for teachers is the realisation of the fact that AI in education is not a tool to wriggle themselves and outsource their core responsibilities to an agentic AI but to transform themselves into a positive catalyst for knowledge transfer in a manner that addresses learner heterogeneity adequately. The use of AI tools to handle large classrooms, personalised teaching-learning, adaptive assessments, early-risk identification, etc. is a teacher’s paradise shifting students form the hell of rot learning.
Initiatives like building sovereign AI, indigenous LLMs, full-stack AI value chain, etc. are long-term projects that cannot have a ‘mini-me’ model in every university or higher educational institution, each claiming a lion’s share as their contribution. This approach is short-circuiting their capabilities from real-time possibilities and drifting them towards unrealistic targets. There is a need to understand institutional capabilities and engage in both consumption and contribution. By being an active participant-consumer in national initiatives like BODHAN AI or BHASHINI or SARVAM AI, native foundational models emerge. By building surface level AI applications to address personalised problems that are abundantly manifest, institutions are self-contributors in their own way.
The need of the hour is a balance in thought and action without being enslaved by a complete auto-functional AI or trapped by legacy pedagogical rigidity. Time has come to draw two converging lines at the drawing board, one that treads the path of AI adoption and the other AI creation. It is in this converging pathway lies the solution that is neither technophilic nor technophobic but contemporarily integrative.
vaidhya@sastra.edu