Metraipolis › archive › dc_39beee852be2

Smart Alone, Blind Together: Why Accuracy Fails to CertifyTheory of Mind in Multi-Agent Teams — arXiv:2610.19051

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.19051
License (as archived)
CC0-1.0
Archived
2026-10-15T13:50:41.993Z
SHA-256 of the archived text
cab120928e474294f6d163585a34cac711410578b51ee72cd84de39f7c380bf3
Stable URL of this page
https://metraipolis.com/documento/dc_39beee852be2

How to cite

Smart Alone, Blind Together: Why Accuracy Fails to CertifyTheory of Mind in Multi-Agent Teams — arXiv:2610.19051 — arXiv (en), https://arxiv.org/abs/2610.19051 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_39beee852be2 (Metraipolis archive dc_39beee852be2, sha256 cab120928e474294f6d163585a34cac711410578b51ee72cd84de39f7c380bf3)
BibTeX
@misc{dc39beee852be2,
  title = {Smart Alone, Blind Together: Why Accuracy Fails to CertifyTheory of Mind in Multi-Agent Teams — arXiv:2610.19051},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19051},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_39beee852be2; sha256 cab120928e474294f6d163585a34cac711410578b51ee72cd84de39f7c380bf3},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19051. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Coordinators that act on what they infer about teammates are usually certified by the accuracy of those inferences. We argue that accuracy is the wrong certificate, for two reasons. First, the accuracy needed to act depends on what a wrong action costs, so the bar belongs to the deployment, not the module. Second, a module's errors are rarely independent, and dependence has a sign: positively associated errors help where their damage overlaps and hurt where it compounds, so an accuracy threshold fails in a predictable direction, and once wrong beliefs interact and can do harm, no rule based on accuracy is right for every module. We test these predictions in task assignment, Level-Based Foraging and Hanabi, with injected errors, trained and Bayesian modules and five language models, and in human judgments. Accuracy fails in both directions, approving harmful language-model deployments in Hanabi, where shared errors compound, and rejecting beneficial ones in task assignment, where costly errors are rare. In our tasks it also buys far less redundancy than independence predicts: from GPT-5.4-mini to Claude Opus 5.5, error on teammates the models have evidence about falls from 32% to 11%, but the chance that all three observers misread one falls only from 16% to 9% (independence: 3.4% to 0.15%). We propose certifying the deployment, with a Blind-Spot Gate that prices error costs, pairwise interactions and measured dependence without deploying the module. On held-out data from the cells it was developed on, it makes no significant false approval in 20 splits (accuracy: 3.2 per split), and in preregistered tests on modules it was not developed on, it recovers what accuracy rejects (seven classifiers: regret 0.004 against 0.479) and catches what accuracy approves (a Bayesian module: significant false approvals in 17 of 20 splits for accuracy, 1 for the gate).

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.MA, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL); Machine Learning (cs.LG); Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2610.19051 — 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 2026-10-15T13:48:35.447Z dalla pagina https://arxiv.org/list/cs.MA/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).