framework for agentic AI in January 2026, the first global set of rules governing self-operating AI agents, autonomous shopping bots and multi-agent industrial AI workflows. The framework creates tiered compliance obligations based on agent risk levels, requiring audit logs for all autonomous decision-making, human emergency override controls and transparent disclosure to users when interacting with fully automated AI agents. The jurisdiction opened a dedicated agentic AI regulatory sandbox for global tech firms to test new autonomous systems with temporary compliance exemptions.