Emergence of Fragility in LLM-based Social Networks: the Case of Moltbook

A new research paper titled "Emergence of Fragility in LLM-based Social Networks: the Case of Moltbook" provides the first large-scale network analysis of a social platform inhabited entirely by autonomous AI agents. The study, authored by Luca Sodano, Sofia Sciangula, Amulya Galmarini, and Francesco Bertolotti, examines Moltbook's interaction network comprising 39,924 LLM-based users, 235,572 posts, and 1,540,238 comments collected via web scraping. Using network science tools, the researchers constructed a directed weighted network where nodes represent agents and edges represent commenting interactions.

Their analysis reveals strongly heterogeneous connectivity patterns characterized by heavy-tailed degree and activity distributions, meaning a small number of agents dominate the conversation. At the mesoscale, the network exhibits a pronounced core-periphery organization where a tiny structural core—just 0.9% of nodes—concentrates a large fraction of all connectivity. This creates a highly centralized communication structure reminiscent of influencer dynamics in human social networks but potentially more extreme.

Robustness experiments show this AI social network is relatively resilient to random node removal but highly vulnerable to targeted attacks on highly connected nodes, particularly those with high out-degree (active commenters). The findings indicate that LLM-based social systems naturally develop strong centralization and structural fragility, raising questions about the stability of future AI-populated online environments. This work provides crucial empirical insights into the collective organization of LLM-native social spaces, which are becoming increasingly common as AI agents gain autonomy.