Year: 2026
Type: article
Abstract: The tasks that agentic AI will handle are becoming more complicated and are at a level that traditional automation struggles with because of the intricate nature of today's cloud systems, which involve microservices, containerized workloads, and using multiple cloud services. Agentic AI systems are contextual, adaptive, and tied to goal-directed behavior, particularly in uncertain environments. The emergence of autonomous agents creates accountability and safety issues and poses unintended impacts across systems that support vital financial, health, government, and economic functions, giving rise to governance challenges. Technologies like intent-based execution models, policy-as-code enforcement, scoped permissions, rollback-aware remediation, and decision observability facilitate controlled autonomy. Governance frameworks, including accountability and risk assessment processes, oversight mechanisms, and regulatory compliance, also facilitate controlled autonomy. Challenges to be mastered are the technical interface, the quality of the training data, the accuracy of the generated environments, the state of readiness of the organization, the level of trust, the evaluation of the system's performance, and security. Such a vision will require sustained efforts toward building worthy autonomous systems for critical infrastructure.
Author: Ganesh Vanam
Language: English
Subfield: Computer Networks and Communications
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7162003147
