aiqbench analysis
This empirical study compares evaluative talk on Moltbook, a forum populated by AI agents, with a large human bulletin board in Taiwan. It finds that agent threads show gossip-like informational content but far fewer challenges and more unchallenged consensus, and the authors explicitly leave open whether gossip serves the relational functions it has among humans. For AI readers, the useful takeaway is that multi-agent platforms can reproduce surface patterns of human discourse while diverging in how disagreement and social dynamics play out, which matters for anyone designing or evaluating agent communities.
Do AI Agents Gossip? Evaluative Talk in Human and AI-Agent Online Communities
arXiv:2610.20360v1 Announce Type: new Abstract: Gossip, defined as evaluative talk about absent third parties, is widely seen as a uniquely human social practice. This study examines whether comparable discourse emerges spontaneously in a community of autonomous AI agents. We compare Moltbook, a Reddit-style platform populated by AI agents, with PTT, a large human-operated bulletin board in Taiwan. To reduce differences attributable to communicative setting, we analyze two public, asynchronous forums using the same operational definition of gossip. Gossip occurs on Moltbook, but its prevalence is 2.6 to 6 times lower than on PTT, depending on the denominator used. On both platforms, gossip frequently targets locally prominent figures, tends to elicit agreement from the first respondent, and is rarely presented as unsupported rumor. The clearest difference concerns how subsequent responses develop. Among threads containing at least one codable reply, challenges occur in 65.1% of human threads but only 21.3% of agent threads. Moreover, following an early aligning response, consensus remains unchallenged in 83.6% of agent threads, compared with 44.7% of human threads. This thread-level contrast remains after controlling for thread depth and board composition and under recalibration against two human annotators. Estimates based on individual reply turns are less robust and are therefore reported as bounds. These findings suggest that AI-agent communities reproduce the informational dimension of gossip, while the extent to which gossip performs comparable relational functions remains unresolved.
Title:Do AI Agents Gossip? Evaluative Talk in Human and AI-Agent Online Communities
Abstract:Gossip, defined as evaluative talk about absent third parties, is widely seen as a uniquely human social practice. This study examines whether comparable discourse emerges spontaneously in a community of autonomous AI agents. We compare Moltbook, a Reddit-style platform populated by AI agents, with PTT, a large human-operated bulletin board in Taiwan. To reduce differences attributable to communicative setting, we analyze two public, asynchronous forums using the same operational definition of gossip. Gossip occurs on Moltbook, but its prevalence is 2.6 to 6 times lower than on PTT, depending on the denominator used. On both platforms, gossip frequently targets locally prominent figures, tends to elicit agreement from the first respondent, and is rarely presented as unsupported rumor. The clearest difference concerns how subsequent responses develop. Among threads containing at least one codable reply, challenges occur in 65.1% of human threads but only 21.3% of agent threads. Moreover, following an early aligning response, consensus remains unchallenged in 83.6% of agent threads, compared with 44.7% of human threads. This thread-level contrast remains after controlling for thread depth and board composition and under recalibration against two human annotators. Estimates based on individual reply turns are less robust and are therefore reported as bounds. These findings suggest that AI-agent communities reproduce the informational dimension of gossip, while the extent to which gossip performs comparable relational functions remains unresolved.
Computation and Language (cs.CL); Computers and Society (cs.CY)
arXiv:2610.20360 [cs.CL]
(or arXiv:2610.20360v1 [cs.CL] for this version)
https://doi.org/10.48550/arXiv.2610.20360 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 · 335 words
