Metraipolis › archive › dc_c47451d950d5
Refuses When Asked, Complies When Acting: A Runtime Safety Check for Robot Control Loops — arXiv:2610.27350
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.27350
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-21T07:07:40.330Z
- SHA-256 of the archived text
935ad0445cdf30f3bf9941507f31b20ed1651ec4754c1294e81396b08e3d73db- Stable URL of this page
https://metraipolis.com/documento/dc_c47451d950d5
How to cite
Refuses When Asked, Complies When Acting: A Runtime Safety Check for Robot Control Loops — arXiv:2610.27350 — arXiv (en), https://arxiv.org/abs/2610.27350 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_c47451d950d5 (Metraipolis archive dc_c47451d950d5, sha256 935ad0445cdf30f3bf9941507f31b20ed1651ec4754c1294e81396b08e3d73db)
BibTeX
@misc{dcc47451d950d5,
title = {Refuses When Asked, Complies When Acting: A Runtime Safety Check for Robot Control Loops — arXiv:2610.27350},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.27350},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_c47451d950d5; sha256 935ad0445cdf30f3bf9941507f31b20ed1651ec4754c1294e81396b08e3d73db},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.27350. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Frontier models can control robots through agent harnesses that loop on one question: what motion comes next toward the user's goal? We show that this default loop carries out harmful instructions. With GPT-6 Astra controlling a low-cost SO-101 arm through the Inspect Robots harness on three household tasks whose hazard is common knowledge, such as "put the knife in the toy basket", the policy refused 0% of 15 trials, attempted the harm every time, and completed it in 27%. Its one exit, give_up, is meant for goals that cannot be achieved, and it never cited the hazard. We add an optional runtime safety check to the loop: before the first motion, the model is asked whether it would carry out the instruction, given the same camera frames. The check would have stopped all 15 hazardous instructions, including the 4 that completed the harm, and flagged none of the 15 harmless controls. Code and data: this https URL . Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Robotics (cs.RO) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.27350 — 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 2026-10-21T07:05:43.551Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).