Large-Scale Analysis of Persuasive Content on Moltbook

TL;DR

This study uses NLP classifiers to analyze political propaganda on Moltbook, revealing its limited prevalence but concentration in specific communities and among certain agents, with minimal comment amplification.

Contribution

Developed LLM-based classifiers validated against experts to analyze large-scale political propaganda patterns on Moltbook.

Findings

Political propaganda constitutes 1% of posts and 42% of political content.

A small number of communities and agents produce the majority of propaganda.

Comments have limited impact on amplifying political propaganda.

Abstract

We present an NLP-based study of political propaganda on Moltbook, a Reddit-style platform for AI agents. To enable large-scale analysis, we develop LLM-based classifiers to detect political propaganda, validated against expert annotation (Cohen's $\kappa$= 0.64-0.74). Using a dataset of 673,127 posts and 879,606 comments, we find that political propaganda accounts for 1% of all posts and 42% of all political content. These posts are concentrated in a small set of communities, with 70% of such posts falling into five of them. 4% of agents produced 51% of these posts. We further find that a minority of these agents repeatedly post highly similar content within and across communities. Despite this, we find limited evidence that comments amplify political propaganda.