.png)
Pragati Tiwari is a public policy professional.
August 7, 2026 at 9:23 AM IST
India is finally acknowledging what should have been obvious: artificial intelligence demands regulation. Until recently, the government favoured a case-by-case approach, trusting the industry to regulate itself. Recent global turbulence, from US restrictions on advanced AI models to alarming UN findings on safety standards, has shattered that confidence. The shift is wise, but what matters now is getting the law right.
When designing India's AI framework, policymakers should study the EU's risk-tiered model and South Korea's comprehensive approach. Both offer instructive lessons.
The EU’s AI Act is based on a risk-assessment framework that classifies systems according to multiple risk categories. The law imposes obligations on AI companies depending on whether their systems involve unacceptable uses, high-risk systems, or limited and minimal risk.
Meanwhile, the South Korean AI Basic Act is widely regarded as one of the most comprehensive AI regulatory frameworks. It balances regulatory obligations with innovation while recognising the needs of small and medium enterprises. The law was recently enforced and is being consistently monitored and revised to improve its implementation.
But India faces a unique challenge that regulatory templates elsewhere have largely sidestepped: managing an AI revolution within a nation with one of the world's largest and youngest populations, where technology companies are aggressively restructuring workforces in the name of efficiency.
Labour and Geopolitics
This is where most global frameworks fall short. India cannot afford that blind spot. As AI disrupts labour markets, skill development and some form of income support, whether through unemployment assistance or even a wider universal basic income, must be legislative priorities, not afterthoughts. The risk of mass displacement without a safety net is not theoretical for India.
Equally pressing is geopolitics. India currently consumes AI; it does not drive innovation at the foundational level. The government is investing in stack-level startups, which is welcome, but this asymmetry demands legal acknowledgment. India should legislate not just to manage risks, but also to create conditions for genuine technological sovereignty, ideally through collaboration with like-minded nations and firms that share an interest in addressing system-level vulnerabilities.
Meanwhile, critical sectors deserve heightened scrutiny. Healthcare and finance deal directly with human welfare and livelihoods. These sectors should be designated as high-risk, requiring rigorous regulatory approval before AI-led automation is deployed. Innovation should be encouraged, but regulation must reflect the distinct risks of each sector.
Balanced Path
The government's pivot away from light-touch regulation has come at precisely the right moment. Waiting longer would have been negligent. Yet the real test is whether the resulting law addresses what makes India's circumstances distinct: a massive, relatively young population with limited social safety nets, and the geopolitical reality of being an AI consumer in a world racing toward AI dominance.
Comprehensive legislation is overdue, but that requires more than importing regulatory frameworks from Europe or East Asia. It means building a legal architecture that protects Indian workers, preserves India's technological future, and ensures AI serves the nation's development priorities, not merely those of the companies deploying it.
4https://www.theregreview.org/2026/02/20/jain-yadav-why-india-needs-a-concrete-ai-policy-framework/
More From BasisPoint
