The evolving landscape of AI agents, also known as operator models, has significant implications for security, authentication, and compliance. These advanced models can interact with user interfaces like humans, unlocking new capabilities but raising concerns about authentication, security, and data protection. To address these challenges, various approaches are being explored, including direct injection, session cookie injection, OAuth and delegated authorization, SSO and Identity Federation, handling MFA and CAPTCHAs, session management and persistence, authorization and scope control, and the Computer-Using Agent (CUA) and new safety paradigms. As these AI agents become more autonomous, ensuring accountability, mitigating risks, and establishing robust security measures are crucial to prevent larger-scale threats and maintain compliance with data protection frameworks like GDPR or CCPA.