MeitY proposes mandatory human checks for agentic AI payments

India's CERT-In proposes mandatory human-in-the-loop controls for agentic AI payments above specified thresholds, with full audit trails, as NPCI develops a Unified Agent Protocol to enable AI agents to make UPI payments. The proposal addresses trust and fraud risks in agentic commerce, following international concerns about autonomous agent spending. If you're building agentic payment infrastructure for India, mandatory human approval above a threshold just became law—your API now needs to route transactions back to a user PIN or dashboard before settlement. That's a rebuild, not a toggle.