India, home to over 245 million enrolled students and the world's largest school education system, has no coherent national policy governing the use of artificial intelligence in education. This commentary argues that the resulting vacuum has produced an "accidental policy," assembled piecemeal through unregulated EdTech product launches, ad hoc teacher decisions, and pragmatic student adoption of generative AI tools. Drawing on the author's empirical research on cognitive-level analysis of Indian school board examinations and on recent evidence that large language models can comfortably clear qualifying thresholds on major Indian entrance examinations such as NEET, the article contends that India's examination architecture is structurally vulnerable to generative AI because it predominantly tests the lower-order cognitive tasks, recall and comprehension, that these systems are optimized to perform. The paper identifies four urgent policy imperatives: fundamental assessment reform, regulation of AI-generated educational content quality, systemic investment in teacher capacity for AI integration, and a national deliberation on the purpose of education in an age of machine cognition. It concludes that the window for deliberate policy-making is narrowing, and that every semester without national guidance is a semester in which accidental policy hardens into institutional default.
Syaamantak Das (2026) studied this question.