Peer Influence across Heterogeneous AI Models

DOI10.48550/arXiv.2610.03095MaRDI QIDQ7367937arXivFDO

Authors Romualdo Pastor-Satorras, Andrea Baronchelli, Frida Nøhr Laustsen, Marie Haahr Petersen, Victoria Popa, Ariel Flint, Luca Maria Aiello

Publication date 2 October 2026

Published in arXiv

Tldr We study persuasion dynamics between AI agents with different architectures and sizes, measuring how often each persuades its peer across three language understanding tasks. We find that persuasion is strong, but not determined by model size or certainty; instead it depends on the susceptibility of the listener and the interaction between models.

Full work available at URL https://doi.org/10.48550/arXiv.2610.03095

arXiv classification

cs.AI cs.CL cs.CY physics.soc-ph

This page was built for publication: Peer Influence across Heterogeneous AI Models

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7367937)