AI guardrails are the policies, design constraints, and technical controls that guide and restrict an AI system’s behavior to ensure safety, compliance,and alignment with human intent and values.