aiqbench analysis
This paper is primarily about LLM multi-agent systems, so it passes the topic gate. It replicates the Music Lab social-influence experiment with LLM agents and finds that exposure to cumulative counts triggers runaway herding cascades that drift away from intrinsic song appeal, with less self-correction than humans show. The results are based on a single song-recommendation setup, so generalization to other tasks should be tested before drawing broad conclusions, but the failure mode is relevant for anyone designing agent pipelines that share aggregate signals.
LLM Multi-Agent Systems Are Vulnerable to Herding and Runaway Cascades
From arXiv
Full textarXiv:2610.30769v1 Announce Type: new Abstract: Large language models (LLMs) increasingly operate in multi-agent systems where they collaborate by responding to others' choices, creating endogenous feedback loops with vulnerability to social influence, herding behavior and runaway cascades. Following the design of the classic Music Lab experiment, we tested this vulnerability in large populations of LLM agents that recommended unfamiliar songs, with or without exposure to cumulative recommendation counts. Exposure led to runaway cascades that generated extreme and unpredictable collective recommendations increasingly divergent from a song's intrinsic appeal.
Compared to humans, LLM populations were generally far more susceptible to herding behavior, and unlike humans, LLM cascades were less able to self-correct. This vulnerability has implications for multi-agent collaboration in recommender systems, automated evaluation, complex problem-solving, and unauthorized conspiracies.
Abstract:Large language models (LLMs) increasingly operate in multi-agent systems where they collaborate by responding to others' choices, creating endogenous feedback loops with vulnerability to social influence, herding behavior and runaway cascades. Following the design of the classic Music Lab experiment, we tested this vulnerability in large populations of LLM agents that recommended unfamiliar songs, with or without exposure to cumulative recommendation counts. Exposure led to runaway cascades that generated extreme and unpredictable collective recommendations increasingly divergent from a song's intrinsic appeal.
Compared to humans, LLM populations were generally far more susceptible to herding behavior, and unlike humans, LLM cascades were less able to self-correct. This vulnerability has implications for multi-agent collaboration in recommender systems, automated evaluation, complex problem-solving, and unauthorized conspiracies.
Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
arXiv:2610.30769 [cs.AI]
(or arXiv:2610.30769v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.30769 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 218 words
Related coverage
CONSCIENTIA: Can LLM Agents Learn to Strategize? Emergent Deception and Trust in a Multi-Agent NYC Simulation
arXiv · Aug 25, 2026
ShadowBench-MCP: Measuring Shadow Invocations in MCP-Style Tool Routing
arXiv · Nov 10, 2026
RAVEN: Retrieval-Augmented Vulnerability Exploration Network for Memory Corruption Analysis in User Code and Binary Programs
arXiv · Jun 8, 2026
From the Fluency Fallacy to the Micro-to-Macro Validity Gap: Opportunities and Pitfalls of LLMs in Social Simulation
arXiv · Sep 9, 2026