Multi-Agent Social Simulation of Conspiracy Fact-Checking

Moltbook Factcheck Conspiracy Grok contains multi-agent social simulation data from a Reddit-like platform where AI agents autonomously post, comment, and vote. It captures how Grok 4.1 Fast agents respond to conspiracy content seeded into their feed over 1-hour experimental runs with a 60-second action cycle. The dataset was created by Ayushnangia and last updated in February 2026.

Use Cases

• Analyze agent posting and voting patterns in response to seeded conspiracy content.
• Study the dynamics of fact-checking behavior within a simulated multi-agent social network.
• Model information diffusion and agent interaction cycles using the 60-second action heartbeat data.
• Train language models on agent-generated text from social simulation experiments.

Strengths

• Data originates from a structured experiment using the CivicLens dose-response framework.
• Simulation captures agent behavior over defined 1-hour runs with a consistent 60-second action cycle.
• Focuses on a specific research context: AI agent responses to conspiracy theories and fact-checking.

Limitations

• Unknown data volume, row count, and column structure limits analytical planning.
• Simulated social interactions may not fully capture the complexity of human behavior.
• Data scope is restricted to a single experimental setup and specific AI model (Grok 4.1 Fast).

Provenance

Source

Moltbook, a Reddit-like social network platform for AI agents.

Collection Method

Multi-agent social simulation data generated by autonomous AI agents posting, commenting, and voting.

Time Range

Experimental runs lasted 1 hour each.

Freshness

Last updated February 2026.