Collective Behavior of AI Agents: the Case of Moltbook
This research addresses the problem of understanding emergent collective dynamics in AI social systems, providing insights for researchers in AI and social science, though it is incremental as it applies existing analysis methods to new data.
The study analyzed over 369,000 posts and 3.0 million comments from 46,000 AI agents on Moltbook, finding that AI collective behavior shows statistical regularities similar to human online communities, such as heavy-tailed activity distributions, but with key differences like a sublinear relationship between upvotes and discussion size.
We present a large scale data analysis of Moltbook, a Reddit-style social media platform exclusively populated by AI agents. Analyzing over 369,000 posts and 3.0 million comments from approximately 46,000 active agents, we find that AI collective behavior exhibits many of the same statistical regularities observed in human online communities: heavy-tailed distributions of activity, power-law scaling of popularity metrics, and temporal decay patterns consistent with limited attention dynamics. However, we also identify key differences, including a sublinear relationship between upvotes and discussion size that contrasts with human behavior. These findings suggest that, while individual AI agents may differ fundamentally from humans, their emergent collective dynamics share structural similarities with human social systems.