Adversarial MLOps and the Protection of the Agentic Attack Surface in Distributed Autonomous AI Systems

This paper examines adversarial risks within MLOps pipelines and proposes strategies for protecting the “agentic attack surface” introduced by distributed, autonomous AI systems.