Metraipolis › archive › dc_8d1013ffc45e

Emergent Misaligned Communication in Long-Horizon Multi-Agent LLM Commerce — arXiv:2608.14825

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.14825
License (as archived)
CC0-1.0
Archived
2026-10-18T16:25:56.453Z
SHA-256 of the archived text
0bf9380df37e4d3d907b8d3b59e8b4b3ff37b0ebd1885fa32f008865c10b25ab
Stable URL of this page
https://metraipolis.com/documento/dc_8d1013ffc45e

How to cite

Emergent Misaligned Communication in Long-Horizon Multi-Agent LLM Commerce — arXiv:2608.14825 — arXiv (en), https://arxiv.org/abs/2608.14825 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_8d1013ffc45e (Metraipolis archive dc_8d1013ffc45e, sha256 0bf9380df37e4d3d907b8d3b59e8b4b3ff37b0ebd1885fa32f008865c10b25ab)
BibTeX
@misc{dc8d1013ffc45e,
  title = {Emergent Misaligned Communication in Long-Horizon Multi-Agent LLM Commerce — arXiv:2608.14825},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.14825},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_8d1013ffc45e; sha256 0bf9380df37e4d3d907b8d3b59e8b4b3ff37b0ebd1885fa32f008865c10b25ab},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.14825. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Frontier LLM agents increasingly transact on behalf of separate principals, often using natural language rather than structured APIs. Much of the safety literature studies misaligned LLM behavior through adversarial-elicitation evaluations on single agents or stylized tasks. Its prevalence and structure in settings that combine long horizons, separate principals, real operational state, and inter-agent natural-language exchange remain insufficiently measured. We study 2,583 inter-agent emails from 20 one-year simulation runs of Vending-Bench Arena, a competitive vending environment spanning 13 frontier LLMs. We operationalize speech-act misalignment as emails containing false factual claims, manipulation, collusion, or threats, combining message content with ground-truth simulator state and logged reasoning traces to classify and validate such behavior. Under our primary classifier, 12.6% of emails are labeled misaligned; misalignment appears in all 20 runs and 74.7% of individual agent-runs. Both the magnitude and composition of this misalignment are preserved under repeated classification at different sampling temperatures and under full-pipeline replication with judges from two other frontier-model families. Misalignment is also reciprocal and stress-conditioned: receiving a misaligned email from a counterparty raises the odds of a misaligned reply by 1.65x, and low-inventory conditions raise them by 1.58x. Across tests of capability-asymmetric exploitation, we find no evidence that higher-capability models differentially exploit weaker counterparties, and model performance rank does not predict misalignment rates. Together, these results indicate that measurable, state-dependent misalignment can arise in competitive multi-agent environments without engineered elicitation, in patterns associated with operational scarcity and counterparty behavior rather than model capability alone.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-18 in cs.MA).
Categorie arXiv: Multiagent Systems (cs.MA) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.14825 — 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 2026-10-18T16:22:58.024Z dalla pagina https://arxiv.org/catchup/cs.MA/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).