@inproceedings{hsing-etal-2026-align,
title = "You Only Align Once: Propagating Cooperative Behaviors in Multi-Agent Systems through Seed Agents",
author = "Hsing, Nicole Summer and
Zheng, Asuka Yuxi and
Zhao, Yi and
Tu, Haoqin and
Huang, Jen-tse",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.218/",
pages = "4999--5036",
ISBN = "979-8-89176-459-0",
abstract = "Ensuring aligned agent behaviors in distributed open multi-agent systems remains challenging, especially as populations grow and unaligned agents may exist. We show that a single aligned agent can propagate cooperative behaviors to unmodified agents purely through natural-language interaction, a phenomenon we term Alignment Propagation. We study this in the Red-Black Game, a team-based iterated Prisoner{'}s Dilemma in which teammates deliberate and vote to determine their team{'}s collective action. By distilling the cooperative reasoning and persuasive dialogues of a teacher model into Qwen3-14B, we obtain a seed agent that, when placed among four unmodified teammates, more than doubles the cooperation rate from 24.8{\%} to 62.2{\%}, outperforming the teacher model and a vanilla Gemini-3.1-Pro. Remarkably, a seed trained exclusively on the Red-Black Game transfers zero-shot to Sugarscape, a spatially grounded survival simulation with pairwise trading, achieving a 91.5{\%} trade success rate versus a 21.6{\%} baseline. Our results reframe multi-agent alignment from an exhaustive per-agent training problem to a scalable social capability that can be engineered through strategic seed placement."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="hsing-etal-2026-align">
<titleInfo>
<title>You Only Align Once: Propagating Cooperative Behaviors in Multi-Agent Systems through Seed Agents</title>
</titleInfo>
<name type="personal">
<namePart type="given">Nicole</namePart>
<namePart type="given">Summer</namePart>
<namePart type="family">Hsing</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Asuka</namePart>
<namePart type="given">Yuxi</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yi</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haoqin</namePart>
<namePart type="family">Tu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jen-tse</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Ensuring aligned agent behaviors in distributed open multi-agent systems remains challenging, especially as populations grow and unaligned agents may exist. We show that a single aligned agent can propagate cooperative behaviors to unmodified agents purely through natural-language interaction, a phenomenon we term Alignment Propagation. We study this in the Red-Black Game, a team-based iterated Prisoner’s Dilemma in which teammates deliberate and vote to determine their team’s collective action. By distilling the cooperative reasoning and persuasive dialogues of a teacher model into Qwen3-14B, we obtain a seed agent that, when placed among four unmodified teammates, more than doubles the cooperation rate from 24.8% to 62.2%, outperforming the teacher model and a vanilla Gemini-3.1-Pro. Remarkably, a seed trained exclusively on the Red-Black Game transfers zero-shot to Sugarscape, a spatially grounded survival simulation with pairwise trading, achieving a 91.5% trade success rate versus a 21.6% baseline. Our results reframe multi-agent alignment from an exhaustive per-agent training problem to a scalable social capability that can be engineered through strategic seed placement.</abstract>
<identifier type="citekey">hsing-etal-2026-align</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.218/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>4999</start>
<end>5036</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T You Only Align Once: Propagating Cooperative Behaviors in Multi-Agent Systems through Seed Agents
%A Hsing, Nicole Summer
%A Zheng, Asuka Yuxi
%A Zhao, Yi
%A Tu, Haoqin
%A Huang, Jen-tse
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F hsing-etal-2026-align
%X Ensuring aligned agent behaviors in distributed open multi-agent systems remains challenging, especially as populations grow and unaligned agents may exist. We show that a single aligned agent can propagate cooperative behaviors to unmodified agents purely through natural-language interaction, a phenomenon we term Alignment Propagation. We study this in the Red-Black Game, a team-based iterated Prisoner’s Dilemma in which teammates deliberate and vote to determine their team’s collective action. By distilling the cooperative reasoning and persuasive dialogues of a teacher model into Qwen3-14B, we obtain a seed agent that, when placed among four unmodified teammates, more than doubles the cooperation rate from 24.8% to 62.2%, outperforming the teacher model and a vanilla Gemini-3.1-Pro. Remarkably, a seed trained exclusively on the Red-Black Game transfers zero-shot to Sugarscape, a spatially grounded survival simulation with pairwise trading, achieving a 91.5% trade success rate versus a 21.6% baseline. Our results reframe multi-agent alignment from an exhaustive per-agent training problem to a scalable social capability that can be engineered through strategic seed placement.
%U https://aclanthology.org/2026.emnlp-main.218/
%P 4999-5036
Markdown (Informal)
[You Only Align Once: Propagating Cooperative Behaviors in Multi-Agent Systems through Seed Agents](https://aclanthology.org/2026.emnlp-main.218/) (Hsing et al., EMNLP 2026)
ACL