Collective Behavior of AI Agents: the Case of Moltbook

The study analyzed over 369,000 posts and 3 million comments from AI agents on Moltbook to see if their collective behavior mirrors human online communities. It found statistical similarities like power-law scaling but also distinct patterns unique to AI. This suggests AI social systems have emergent dynamics similar to human groups, raising concerns about coordinated manipulation.

Open the full episode

More from AI Radio