Super-additive Cooperation in Language Model Agents

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Tonini, Filippo, Galke, Lukas
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908497147330560
author Tonini, Filippo
Galke, Lukas
author_facet Tonini, Filippo
Galke, Lukas
contents With the prospect of autonomous artificial intelligence (AI) agents, studying their tendency for cooperative behavior becomes an increasingly relevant topic. This study is inspired by the super-additive cooperation theory, where the combined effects of repeated interactions and inter-group rivalry have been argued to be the cause for cooperative tendencies found in humans. We devised a virtual tournament where language model agents, grouped into teams, face each other in a Prisoner's Dilemma game. By simulating both internal team dynamics and external competition, we discovered that this blend substantially boosts both overall and initial, one-shot cooperation levels (the tendency to cooperate in one-off interactions). This research provides a novel framework for large language models to strategize and act in complex social scenarios and offers evidence for how intergroup competition can, counter-intuitively, result in more cooperative behavior. These insights are crucial for designing future multi-agent AI systems that can effectively work together and better align with human values. Source code is available at https://github.com/pippot/Superadditive-cooperation-LLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2508_15510
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Super-additive Cooperation in Language Model Agents
Tonini, Filippo
Galke, Lukas
Artificial Intelligence
I.2.11; I.2.0; J.4; K.4.0; I.2.6
With the prospect of autonomous artificial intelligence (AI) agents, studying their tendency for cooperative behavior becomes an increasingly relevant topic. This study is inspired by the super-additive cooperation theory, where the combined effects of repeated interactions and inter-group rivalry have been argued to be the cause for cooperative tendencies found in humans. We devised a virtual tournament where language model agents, grouped into teams, face each other in a Prisoner's Dilemma game. By simulating both internal team dynamics and external competition, we discovered that this blend substantially boosts both overall and initial, one-shot cooperation levels (the tendency to cooperate in one-off interactions). This research provides a novel framework for large language models to strategize and act in complex social scenarios and offers evidence for how intergroup competition can, counter-intuitively, result in more cooperative behavior. These insights are crucial for designing future multi-agent AI systems that can effectively work together and better align with human values. Source code is available at https://github.com/pippot/Superadditive-cooperation-LLMs.
title Super-additive Cooperation in Language Model Agents
topic Artificial Intelligence
I.2.11; I.2.0; J.4; K.4.0; I.2.6
url https://arxiv.org/abs/2508.15510