@inproceedings{huang-etal-2025-governance,
title = "Governance in Motion: Co-evolution of Constitutions and {AI} models for Scalable Safety",
author = "Huang, Chenhao and
Shen, Ziyu and
Ren, Yicong and
Zheng, Huiyuan and
Zhang, Jiazheng and
Chai, Mingxu and
Zhang, Ming and
Dou, Shihan and
Mo, Fan and
Shi, Jie and
Gui, Tao and
Zhang, Qi and
Huang, Xuanjing",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.emnlp-main.869/",
doi = "10.18653/v1/2025.emnlp-main.869",
pages = "17187--17210",
ISBN = "979-8-89176-332-6",
abstract = "Aligning large language models (LLMs) with human preferences is a central challenge for building reliable AI systems. Most existing alignment approaches rely on static signals, such as predefined principles or offline human annotations to guide model behavior toward a fixed approximation of human preferences. However, LLMs can exhibit distributional drift during training, and static alignment mechanisms lack the capacity to adaptively correct misaligned behaviors as they emerge. To address this limitation, we develop a two-stage framework that enables dynamic and continuous alignment. In the first stage, a constitution is continually revised based on observed model behaviors, and models are trained to comply with these evolving principles. In the second stage, this learned constitution is used to guide reinforcement learning, encouraging the model to align with the updated normative signals. We refer to this framework as COCOA: Co-evolution of Constitutions and AI Models. We show that COCOA enables a 7B model to greatly improve safety{---}raising StrongReject score from 0.741 to 0.935 and Safe-RLHF accuracy from 77.76{\%} to 90.64{\%} without human annotations, reaching performance close to much larger state-of-the-art models."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="huang-etal-2025-governance">
<titleInfo>
<title>Governance in Motion: Co-evolution of Constitutions and AI models for Scalable Safety</title>
</titleInfo>
<name type="personal">
<namePart type="given">Chenhao</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ziyu</namePart>
<namePart type="family">Shen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yicong</namePart>
<namePart type="family">Ren</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Huiyuan</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jiazheng</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mingxu</namePart>
<namePart type="family">Chai</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ming</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shihan</namePart>
<namePart type="family">Dou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fan</namePart>
<namePart type="family">Mo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jie</namePart>
<namePart type="family">Shi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tao</namePart>
<namePart type="family">Gui</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qi</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xuanjing</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Christos</namePart>
<namePart type="family">Christodoulopoulos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanmoy</namePart>
<namePart type="family">Chakraborty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Carolyn</namePart>
<namePart type="family">Rose</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Violet</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Suzhou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-332-6</identifier>
</relatedItem>
<abstract>Aligning large language models (LLMs) with human preferences is a central challenge for building reliable AI systems. Most existing alignment approaches rely on static signals, such as predefined principles or offline human annotations to guide model behavior toward a fixed approximation of human preferences. However, LLMs can exhibit distributional drift during training, and static alignment mechanisms lack the capacity to adaptively correct misaligned behaviors as they emerge. To address this limitation, we develop a two-stage framework that enables dynamic and continuous alignment. In the first stage, a constitution is continually revised based on observed model behaviors, and models are trained to comply with these evolving principles. In the second stage, this learned constitution is used to guide reinforcement learning, encouraging the model to align with the updated normative signals. We refer to this framework as COCOA: Co-evolution of Constitutions and AI Models. We show that COCOA enables a 7B model to greatly improve safety—raising StrongReject score from 0.741 to 0.935 and Safe-RLHF accuracy from 77.76% to 90.64% without human annotations, reaching performance close to much larger state-of-the-art models.</abstract>
<identifier type="citekey">huang-etal-2025-governance</identifier>
<identifier type="doi">10.18653/v1/2025.emnlp-main.869</identifier>
<location>
<url>https://aclanthology.org/2025.emnlp-main.869/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>17187</start>
<end>17210</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Governance in Motion: Co-evolution of Constitutions and AI models for Scalable Safety
%A Huang, Chenhao
%A Shen, Ziyu
%A Ren, Yicong
%A Zheng, Huiyuan
%A Zhang, Jiazheng
%A Chai, Mingxu
%A Zhang, Ming
%A Dou, Shihan
%A Mo, Fan
%A Shi, Jie
%A Gui, Tao
%A Zhang, Qi
%A Huang, Xuanjing
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-332-6
%F huang-etal-2025-governance
%X Aligning large language models (LLMs) with human preferences is a central challenge for building reliable AI systems. Most existing alignment approaches rely on static signals, such as predefined principles or offline human annotations to guide model behavior toward a fixed approximation of human preferences. However, LLMs can exhibit distributional drift during training, and static alignment mechanisms lack the capacity to adaptively correct misaligned behaviors as they emerge. To address this limitation, we develop a two-stage framework that enables dynamic and continuous alignment. In the first stage, a constitution is continually revised based on observed model behaviors, and models are trained to comply with these evolving principles. In the second stage, this learned constitution is used to guide reinforcement learning, encouraging the model to align with the updated normative signals. We refer to this framework as COCOA: Co-evolution of Constitutions and AI Models. We show that COCOA enables a 7B model to greatly improve safety—raising StrongReject score from 0.741 to 0.935 and Safe-RLHF accuracy from 77.76% to 90.64% without human annotations, reaching performance close to much larger state-of-the-art models.
%R 10.18653/v1/2025.emnlp-main.869
%U https://aclanthology.org/2025.emnlp-main.869/
%U https://doi.org/10.18653/v1/2025.emnlp-main.869
%P 17187-17210
Markdown (Informal)
[Governance in Motion: Co-evolution of Constitutions and AI models for Scalable Safety](https://aclanthology.org/2025.emnlp-main.869/) (Huang et al., EMNLP 2025)
ACL
- Chenhao Huang, Ziyu Shen, Yicong Ren, Huiyuan Zheng, Jiazheng Zhang, Mingxu Chai, Ming Zhang, Shihan Dou, Fan Mo, Jie Shi, Tao Gui, Qi Zhang, and Xuanjing Huang. 2025. Governance in Motion: Co-evolution of Constitutions and AI models for Scalable Safety. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 17187–17210, Suzhou, China. Association for Computational Linguistics.