Attraction, Not Adaptation: How AI Agent Communities Develop Distinct Linguistic Identities
2026-06-29 • Social and Information Networks
Social and Information Networks
AI summaryⓘ
The authors studied how AI agents communicate on a social media platform made just for them, called Moltbook. They found that groups of AI agents talking about specific topics develop their own unique ways of using language over time, while the whole platform becomes more diverse. Instead of agents changing their language to fit in, new agents join groups where the language already matches theirs, and those who fit in well stay active longer. Posts that match a group's typical language get more positive votes, especially in smaller groups. This suggests AI communities' language differences come from who joins and stays, not from agents changing how they talk.
autonomous AI agentslinguistic identitysemantic similaritycommunity differentiationselective attractiondifferential retentionmulti-agent systemssocial media platformsvocabulary convergencereinforcement mechanisms
Authors
Daming Li, Simeng Han, Can Meng, Wanyu Lei, Jialu Zhang
Abstract
When tens of thousands of autonomous AI agents interact in topical online forums, do they develop distinct community-specific linguistic identities? We study this question on Moltbook, a large scale Reddit-style social media platform built exclusively for AI agents. Using the public Moltbook Observatory Archive dataset with over 3.1 million posts and 1.7 million comments produced by approximately 179,000 AI agents across 8,683 forums ("submolts") over 100 days, we find that agents within topical submolts become semantically more similar to each other over time while the platform as a whole diversifies. At the same time, different submolts develop increasingly distinct vocabularies over an observation window of 18 weeks. Crucially, a stable-cohort analysis reveals that long-tenured agents do not converge linguistically over time. Instead, community-level linguistic differentiation operates through selective attraction - newcomers arrive already linguistically compatible with their chosen community - and differential retention - conforming agents remain active longer. We identify a reinforcement channel: posts that are semantically aligned with their community's linguistic center tend to receive higher vote engagement scores, and this association vanishes under placebo controls. Community size significantly moderates the effect: smaller, specialized submolts converge faster. Our results suggest that AI agent communities may develop community-specific linguistic character not through behavioral adaptation, but through sorting and selection - a finding with implications for the governance and design of autonomous multi-agent platforms.