According to Help Net Security, security measures for AI agent APIs have not kept pace with the deployment of AI agents already operating in production environments. This gap between security infrastructure and active implementations presents a potential risk for organizations utilizing AI agents in their operations.