Metraipolis › archive › dc_2020a075e461
Limits of Statistical Assurance for Embodied and Agentic AI under Adaptive Adversaries — arXiv:2610.18948
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.18948
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-15T14:23:10.575Z
- SHA-256 of the archived text
5ffb3f9b804f780dd5cec036ea004614429a2d17e7c16a9195fb143f39f18980- Stable URL of this page
https://metraipolis.com/documento/dc_2020a075e461
How to cite
Limits of Statistical Assurance for Embodied and Agentic AI under Adaptive Adversaries — arXiv:2610.18948 — arXiv (en), https://arxiv.org/abs/2610.18948 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2020a075e461 (Metraipolis archive dc_2020a075e461, sha256 5ffb3f9b804f780dd5cec036ea004614429a2d17e7c16a9195fb143f39f18980)
BibTeX
@misc{dc2020a075e461,
title = {Limits of Statistical Assurance for Embodied and Agentic AI under Adaptive Adversaries — arXiv:2610.18948},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.18948},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2020a075e461; sha256 5ffb3f9b804f780dd5cec036ea004614429a2d17e7c16a9195fb143f39f18980},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.18948. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
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 probability and $S_0$ the unattacked certification time. The result covers adaptive and privately randomized evaluators. Switching at certification is exactly optimal for a fixed likelihood-ratio test under memoryless, event-increasing interventions without a binding intervention budget. A positive clean-certification requirement forces linear worst-case expected excess when intervention raises event rates over an authorized deployment horizon, even if certification is deferred to the deadline. In simulated RoCo collaboration, switching yields commanded-separation violations in 34 of 36 post-certification episodes. Preregistered randomized AgentDojo trials estimate 1.98 additional security violations per four-task continuation (95% interval [1.67, 2.29]). An adversarial language-model controller infers hidden certification from noisy evaluation receipts and times intervention accordingly. With permanent revocation, a mixture likelihood-ratio monitor bounds cumulative excess above a prespecified reference rate at square-root order, up to logarithmic factors, on every sample path. A single fixed alternative can permit linear expected excess under stationary attacks. Against a controller observing the monitor's rule and statistic, live mixture monitoring reduces violations from 12.00 to 3.81 per 24 assigned task opportunities. Revocation also reduces tasks executed.
Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Cryptography and Security (cs.CR) ; Machine Learning (cs.LG); Machine Learning (stat.ML).
Pagina del lavoro: https://arxiv.org/abs/2610.18948 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T13:53:35.811Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).