Regulating AI in education: India’s enablement-first approach

Education is a high-impact, high-sensitivity sector. Decisions influenced by Artificial Intelligence (AI) such as learning progression, assessment outcomes or access to employment opportunities can have lifelong consequences across stakeholders especially the learners. At the same time, India’s education ecosystem is characterised by varying institutional capacity, learner preparedness, language contexts and levels of digital maturity. Alongside this, are some real risks such as learner data privacy and misuse, AI tutoring that over-personalises and replaces critical human mentorship, models trained on global datasets misaligned with India’s linguistic, cultural, pedagogical diversity and exclusivity from paywalled platforms that widens inequality.

Global policy guidance and empirical research have raised similar concerns. International frameworks on the ethical adoption of AI underline the need for strong human-rights protections and caution against surveillance risks, especially for vulnerable populations¹. Policy analysis has also emphasised privacy, bias and autonomy risks in education use-cases². Academic research has linked excessive screen time to cognitive/behavioral harms³.

Against this backdrop, ‘one-size fits-all’ regulatory models built for homogeneous systems or smaller markets cannot be directly adopted. Instead, India’s response should be enablement-first, light-but-tight approach i.e. encourage experimentation while tightly guarding public interest. This is consistent with the National Education Policy, which promotes technology for personalisation and access while embedding safeguards and with India’s national AI strategy that prioritises foundational infrastructure and locally relevant datasets.

A fit-for-purpose AI framework for education should cover the entire value chain from curriculum design, content creation, formative assessment, tutoring and learning analytics to administrative systems (admissions, credentialing), teacher professional development and labour market linkages. In our view, it revolves around five areas:

Across India, educational institutions are integrating AI in ways that reflect an enablement-oriented philosophy. Some of these initiatives include AI-driven learning analytics to identify student learning gaps while preserving faculty authority; language and speech technologies to expand access across multiple Indian languages; embedding AI into academic programmes with modules on ethics and responsible use. A notable example of a leading institution which allows students to use generative AI tools in exams and assignments but requires them to submit the prompts used. This shifts evaluation from rote memory to prompt quality, reasoning and critical thinking. This reflects a system that wants to create active innovators and not passive consumers.

An enablement-first approach can be operationalised through pragmatic policy instruments, like:

Many of these measures can be implemented through guidelines, funding conditions and institutional mandates without requiring immediate legislation. Over the next 12 to 24 months, India has an opportunity to shape AI governance in education by publishing baseline trust standards for high-risk AI education use cases, piloting state-level regulatory sandboxes and investing in capacity building.

The objective is not to eliminate risk, which is neither feasible nor desirable, but to ensure AI in education remains transparent, accountable and aligned with public purpose. For India, the right path is pragmatic and contextual i.e. build open standards and local datasets, experiment prudently across the whole education value chain and uphold the highest trust standards. Enable innovation however with fairness, accountability and national purpose at the core.