Metraipolis › archive › dc_d2afdb3e3965

When "No Worries" Went Wrong: HRI Lessons for Foundation-Model Robots — arXiv:2610.38400

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.38400
License (as archived)
CC0-1.0
Archived
2026-11-01T02:44:14.041Z
SHA-256 of the archived text
655f4b083b804e16d914bba9c0193a7a0458f4401d17a468261a5fa649302ac0
Stable URL of this page
https://metraipolis.com/documento/dc_d2afdb3e3965

How to cite

When "No Worries" Went Wrong: HRI Lessons for Foundation-Model Robots — arXiv:2610.38400 — arXiv (en), https://arxiv.org/abs/2610.38400 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d2afdb3e3965 (Metraipolis archive dc_d2afdb3e3965, sha256 655f4b083b804e16d914bba9c0193a7a0458f4401d17a468261a5fa649302ac0)
BibTeX
@misc{dcd2afdb3e3965,
  title = {When "No Worries" Went Wrong: HRI Lessons for Foundation-Model Robots — arXiv:2610.38400},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.38400},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d2afdb3e3965; sha256 655f4b083b804e16d914bba9c0193a7a0458f4401d17a468261a5fa649302ac0},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.38400. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

When a person declined our robot's invitation to play, Alice turned her shoulder away and said "No worries, it's all good." Only after a brief pause did she drive off. The words and gestures were plausible but their order changed the message. In another episode, Alice informed a host that his visitor had arrived---while the visitor stood right beside him. Both came from a three-year ROS 2 HRI consortium that integrated large language and vision models as they emerged. Today's foundation models give robots broader capabilities to perceive their surroundings, reason about situations and generate behavior. Yet capability alone does not ensure that a changed situation shapes the action that reaches a person. Evidence may never be acquired, or a purpose or constraint may fail to survive planning and execution. A system may even report Success when an action no longer serves its intended purpose. Drawing on the consortium's engineering experience, we develop five design lessons about attention, knowledge, commitments, embodied timing and action authority. We then introduce TRACE, a lightweight review that changes one interaction premise, follows it to a decision or enforcement boundary, and produces an actionable architecture finding. A bounded analysis of a published LLM-generated behavior-tree system illustrates how an available clarification mechanism can go unused. The engineering lesson is to preserve accountability boundaries, not the computational pipelines that once exposed them.

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.RO, nuovi lavori del giorno).
Categorie arXiv: Robotics (cs.RO).
Pagina del lavoro: https://arxiv.org/abs/2610.38400 — 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 2026-11-01T02:16:58.628Z dalla pagina https://arxiv.org/list/cs.RO/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).