AI Safety Rules Without Chilling Open Models
Risk-based regulation can coexist with open weights if duties sit on deployment, not research.
Model releases, compute thresholds and content provenance are entering Indian policy debate. Copying the EU Act wholesale would misfit a market still building domestic capability.
A better path assigns duties to high-risk uses — credit, hiring, health, welfare — and keeps research and open weights less burdened.
