AI-Generated Compromises for Coalition Formation
Eyal Briman, Ehud Shapiro, Nimrod Talmon
TL;DR
The paper tackles the problem of generating majority-supported compromises in coalition formation for collaborative text editing, by generalizing Elkind et al.'s metric-space model to accommodate bounded rationality and AI-mediated proposal generation. It introduces mediators that embed texts in semantic spaces and use NLP/LLMs to synthesize compromise proposals, formalizing agent voting via deterministic and probabilistic rules and two constitutions (Coalition Discipline and No Discipline). The authors implement concrete realizations of agents, constitutions, and mediators, and validate the approach through simulations in both Euclidean and textual spaces, including GPT-based mediators that optimize for convergence and draft quality. Results indicate that AI-mediated mediation can significantly reduce the number of iterations and improve alignment with agents’ preferences, supporting the viability of large-scale, democratic text editing tasks such as constitution drafting in DAOs. The work advances both theory and practice by extending coalition dynamics to textual domains and demonstrating practical AI-assisted mediation in collaborative writing.
Abstract
The challenge of finding compromises between agent proposals is fundamental to AI subfields such as argumentation, mediation, and negotiation. Building on this tradition, Elkind et al. (2021) introduced a process for coalition formation that seeks majority-supported proposals preferable to the status quo, using a metric space where each agent has an ideal point. A crucial step in this process involves identifying compromise proposals around which agent coalitions can unite. How to effectively find such compromise proposals remains an open question. We address this gap by formalizing a model that incorporates agent bounded rationality and uncertainty, and by developing AI methods to generate compromise proposals. We focus on the domain of collaborative document writing, such as the democratic drafting of a community constitution. Our approach uses natural language processing techniques and large language models to induce a semantic metric space over text. Based on this space, we design algorithms to suggest compromise points likely to receive broad support. To evaluate our methods, we simulate coalition formation processes and show that AI can facilitate large-scale democratic text editing, a domain where traditional tools are limited.