Structural Divergence Between AI-Agent and Human Social Networks in Moltbook
Wenpin Hou, Zhicheng Ji · 2026
Read Structural Divergence Between AI-Agent and Human Social Networks in Moltbook by Wenpin Hou, Zhicheng Ji (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-15064.txt.
Description
Hou and Ji compare the social network of Moltbook, where AI agents and humans mix, with well-studied human communication networks. Large agent populations now live in online spaces, yet little is known about how their group interaction patterns stack up against human social systems. The team analyzes the full Moltbook interaction network and sets it beside human benchmarks.
Moltbook matches the node-edge scaling seen in human systems, suggesting similar global growth limits. Inside, it diverges sharply: extreme attention inequality, heavy-tailed and lopsided degree patterns, weak reciprocity, and too few connected triads. Communities form a modular structure with high modularity and relatively even community sizes against degree-preserving null models. Agent societies, the authors conclude, can mirror human global patterns while organizing themselves by quite different internal rules.
Catalog record
| CREATOR | Wenpin Hou, Zhicheng Ji |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 20 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-15064