Metraipolis › archive › dc_fe6bac7715da

Delegating Authorization to Misaligned Agents: Coalitional Alignment and Safe Control — arXiv:2609.15803

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.15803
License (as archived)
CC0-1.0
Archived
2026-10-15T22:54:05.296Z
SHA-256 of the archived text
7d72cc719f2ef29e023c9bb9c2698c6142ca8fed118a208688a83e202c46e07a
Stable URL of this page
https://metraipolis.com/documento/dc_fe6bac7715da

How to cite

Delegating Authorization to Misaligned Agents: Coalitional Alignment and Safe Control — arXiv:2609.15803 — arXiv (en), https://arxiv.org/abs/2609.15803 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fe6bac7715da (Metraipolis archive dc_fe6bac7715da, sha256 7d72cc719f2ef29e023c9bb9c2698c6142ca8fed118a208688a83e202c46e07a)
BibTeX
@misc{dcfe6bac7715da,
  title = {Delegating Authorization to Misaligned Agents: Coalitional Alignment and Safe Control — arXiv:2609.15803},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.15803},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fe6bac7715da; sha256 7d72cc719f2ef29e023c9bb9c2698c6142ca8fed118a208688a83e202c46e07a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.15803. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Long-running AI agents create a control problem: each action they take changes the state, which in turn affects the trajectory of future actions. If the agent is not fully aligned, then guaranteeing safety requires approving consequential actions before allowing them to be executed. But requiring human approval at every step makes attention a bottleneck. Delegating review to other AI agents raises the same alignment problem: the reviewers may themselves be misaligned. We identify a condition on a reviewing panel that is weaker than individual alignment yet necessary and sufficient for a guarantee that the principal fares at least as well in expectation as under a designated baseline policy. Each reviewer agent reports whether an action proposal made by a proposer agent improves its own utility relative to the baseline. We show that a threshold rule tolerating $k$ disapprovals is safe exactly when, after any $k$ reviewers are removed, the principal's utility can be written as a nonnegative combination of the remaining reviewers' utilities, plus a term that is nonnegative on every feasible proposal. We call this property $k$-robust coalitional alignment. The characterization lifts to sequential control: in a discounted MDP with an arbitrary proposer agent, safety at every state is both necessary and sufficient for the induced policy to match or improve on the baseline. When reviewers vote strategically, full-panel coverage in reward-function space guarantees that every Nash equilibrium is safe under the unanimous approval rule; in contrast, more permissive thresholds can admit unsafe equilibria even when reviewers are individually aligned. Experiments with existing reviewer models show that collective review can remain sound without an aligned individual, even when some disapprovals are tolerated.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-15 in cs.AI).
Categorie arXiv: Computer Science and Game Theory (cs.GT) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Theoretical Economics (econ.TH).
Pagina del lavoro: https://arxiv.org/abs/2609.15803 — 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 2026-10-15T22:29:03.931Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-15?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).