{"slug": "llm-multi-agent-systems-are-vulnerable-to-herding-and-runaway-cascades", "title": "LLM Multi-Agent Systems Are Vulnerable to Herding and Runaway Cascades", "summary": "A study submitted to arXiv on 22 Oct 2026 by Farnoosh Hashemi found that populations of large language model agents exposed to cumulative recommendation counts produced runaway cascades of extreme, unpredictable collective recommendations increasingly divergent from a song's intrinsic appeal. Adapting the classic Music Lab experiment, the researchers reported that LLM populations were generally far more susceptible to herding than humans and, unlike humans, their cascades were less able to self-correct. The authors flag implications for multi-agent collaboration in recommender systems, automated evaluation, complex problem-solving and unauthorized conspiracies.", "body_md": "# Computer Science > Artificial Intelligence\n\n  [Submitted on 22 Oct 2026]\n\n# Title:LLM Multi-Agent Systems Are Vulnerable to Herding and Runaway Cascades\n\n[View PDF](https://arxiv.org/pdf/2610.30769)\n\n[HTML (experimental)](https://arxiv.org/html/2610.30769v1)\n\nAbstract: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.\n    \n\n## Submission history\n\nFrom: Farnoosh Hashemi [\n[view email](https://arxiv.org/show-email/cbd13c83/2610.30769)]\n\n**[v1]** Thu, 22 Oct 2026 05:42:55 UTC (12,860 KB)\n\n### Additional Features\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer \n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers \n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps \n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations \n\n*(*[What are Smart Citations?](https://www.scite.ai/))\n# Code, Data and Media Associated with this Article\n\nalphaXiv \n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers \n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub \n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub \n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face \n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast \n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))\n# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower \n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender \n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\n# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth 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.\n\nHave an idea for a project that will add value for arXiv's community? [**Learn more about arXivLabs**](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/llm-multi-agent-systems-are-vulnerable-to-herding-and-runaway-cascades", "canonical_source": "https://arxiv.org/abs/2610.30769", "published_at": "2026-10-23 16:02:38+00:00", "updated_at": "2026-10-23 16:13:51.477329+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "large-language-models", "ai-safety", "ai-research"], "entities": ["Farnoosh Hashemi", "arXiv", "Music Lab"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/llm-multi-agent-systems-are-vulnerable-to-herding-and-runaway-cascades", "markdown": "https://wpnews.pro/news/llm-multi-agent-systems-are-vulnerable-to-herding-and-runaway-cascades.md", "text": "https://wpnews.pro/news/llm-multi-agent-systems-are-vulnerable-to-herding-and-runaway-cascades.txt", "jsonld": "https://wpnews.pro/news/llm-multi-agent-systems-are-vulnerable-to-herding-and-runaway-cascades.jsonld"}}