Limits of Statistical Assurance for Embodied and Agentic AI under Adaptive Adversaries

본문 미리보기

arXiv:2610.18948v1 Announce Type: new Abstract: Finite safety assessments cannot exclude an adversary that preserves unattacked behavior until certification and activates afterward. We characterize this limitation for embodied and tool-use AI. An adversary that matches the complete observation law until observable, irrevocable certification and retains intervention access afterward attains $v\mathbb{E}[(n-S_0)_+]$ expected post-certification events over $n$ trials, where $v$ is the active event