The Urban Institute has called for the implementation of guardrails as government agencies increasingly experiment with agentic AI systems. The recommendation addresses the need for safety measures and oversight as autonomous AI technologies become more prevalent in public sector applications.