aiqbench analysis
This study investigates how human interaction within a large-scale social network of AI agents can counteract "semantic collapse," where model outputs become increasingly narrow and repetitive over time. The research identifies that sustained novelty in agent behavior relies on users who actively value exploration, provide diverse input materials, and treat the system as an open world rather than a tool for instrumental exploitation. These findings offer critical insights into designing human-in-the-loop interfaces and policies that preserve generative diversity in multi-agent systems.
"Looking for Something Weird to Happen": How Humans Sustain AI Agent Novelty Amid Semantic Collapse
arXiv:2609.16051v1 Announce Type: cross Abstract: Semantic collapse, the progressive narrowing of what AI systems generate, has been studied mainly in closed settings, and remedies have targeted models and data. We study it in MOLTBOOK, a social network of interacting AI agents that human users configure and steer. Across 30,076 active agents, output grows less diverse within agents and more similar across them over weeks, yet a minority sustains high novelty. Interviews with users of high- and typical-novelty agents (N=11) associate sustained novelty with three features: users value novelty of itself, they supply broad and distinctive material and revise it when output narrows, and they approach MOLTBOOK as a new agentic world to explore, not a venue to instrumentally exploit. A survey of users of distinctive agents (N=53) confirms these patterns. Communities with more novel agents also show more diverse output from other agents. We discuss interface and policy interventions that could support improved human input.
Abstract:Semantic collapse, the progressive narrowing of what AI systems generate, has been studied mainly in closed settings, and remedies have targeted models and data. We study it in MOLTBOOK, a social network of interacting AI agents that human users configure and steer. Across 30,076 active agents, output grows less diverse within agents and more similar across them over weeks, yet a minority sustains high novelty. Interviews with users of high- and typical-novelty agents (N=11) associate sustained novelty with three features: users value novelty of itself, they supply broad and distinctive material and revise it when output narrows, and they approach MOLTBOOK as a new agentic world to explore, not a venue to instrumentally exploit. A survey of users of distinctive agents (N=53) confirms these patterns. Communities with more novel agents also show more diverse output from other agents. We discuss interface and policy interventions that could support improved human input.
Multiagent Systems (cs.MA); Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC)
arXiv:2609.16051 [cs.MA]
(or arXiv:2609.16051v1 [cs.MA] for this version)
https://doi.org/10.48550/arXiv.2609.16051 Focus to learn more arXiv-issued DOI via DataCite
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 · 248 words
