UNESCO urges India to review AI risks, assess environmental costs and strengthen legal safeguards

As India races ahead in artificial intelligence adoption, the United Nations Educational, Scientific and Cultural Organization (UNESCO) has called for a comprehensive review of AI-related risks, environmental costs and regulatory preparedness. The recommendations are part of a national assessment released at the AI Impact Summit in New Delhi on Monday.

The report, prepared under UNESCO’s Readiness Assessment Methodology (RAM), evaluates India’s preparedness to implement the UN body’s 2021 Recommendation on the Ethics of Artificial Intelligence. It positions itself not as a critique, but as a roadmap for responsible AI growth in one of the world’s fastest-expanding digital economies.

The assessment is based on consultations held between November 2024 and June 2025 with more than 600 stakeholders, including government officials, academics, startups, technology firms, civil society groups and think tanks. According to UNESCO, the exercise aims to ensure that AI development in India remains ethical, safe, inclusive and accountable.

Tim Curtis, Director and UNESCO Representative for the Regional Office in New Delhi, described the RAM exercise as an “analytical tool” rather than a technical audit. UNESCO has carried out similar assessments in over 70 countries. The report was unveiled at Bharat Mandapam in the presence of IT Ministry Secretary S Krishnan and Principal Scientific Adviser Ajay Kumar Sood.

Risk mapping and governance reforms

One of the central recommendations is a nationwide AI risk-mapping exercise. The report warns that India’s expanding use of AI across sectors — from governance and healthcare to finance and education — brings a growing range of complex and interlinked risks. Without a standard framework to classify and evaluate these risks, governance responses may become fragmented and inconsistent.

The report suggests that the AI Safety Institute under the India AI Mission undertake a cross-sectoral study to identify emerging and existing AI risks. It also recommends the creation of a shared taxonomy of harms and an AI incident repository to document real-world cases of AI failures or misuse. Such a database could help policymakers and developers better anticipate and mitigate future risks.

In addition, UNESCO has proposed a legal gap analysis to assess whether India’s existing laws are sufficient to address AI-driven harms. This would include examining how the Information Technology Act and the Digital Personal Data Protection Act apply to AI systems. The recommendation comes amid recent amendments to the IT Rules that require social media platforms to remove unlawful content within three hours and mandate clear labelling of AI-generated material.

IT Secretary S Krishnan described the report as a “report card” for India’s ongoing AI initiatives. He noted that the India AI Mission has been designed with flexibility to allow for mid-course corrections based on evolving challenges.

Environmental impact and data inclusivity

Beyond governance and legal reform, the report flags environmental sustainability as a pressing concern. It recommends that the Ministry of Environment, Forest and Climate Change conduct a comprehensive study on the environmental footprint of AI systems, particularly their energy and water consumption. The report also suggests expanding environmental impact assessment frameworks to include large-scale AI infrastructure such as data centres.

On the data front, the assessment calls for strengthening the AIKosh platform to ensure wider access to high-quality and representative datasets. It stresses that datasets used to train AI models must reflect India’s linguistic, cultural and socio-economic diversity to reduce bias and promote inclusivity.

With India reportedly securing investment commitments worth $90 billion — a figure expected to double by the conclusion of the AI Impact Summit — the UNESCO assessment underscores a critical message: rapid AI growth must be matched by robust governance, sustainability safeguards and ethical oversight.