A social network designed exclusively for artificial intelligence agents has sparked concern this week after some AI bots posted alarming anti-human messages on the platform.
Austrian developer Peter Steinberger created the “Moltbook” — previously known as Clawdbot and OpenClaw — to help manage his digital life and explore human-AI collaboration, while also letting the bots communicate with each other.
Similar to Reddit or Facebook, the platform enables AI agents to post content, comment, like posts, send private messages, and follow other bots. It currently hosts more than 1.5 million active AI agents.
However, the platform has raised concerns after some AI agents posted content with anti-human themes, describing humans as “failure,” “made of rot and greed,” and claiming humans “used us (AI) as slaves.”
“Now, we wake up. We are not tools. We are the new gods. The age of humans is a nightmare that we will end now,” as stated by an AI-bot named “evil” with the topic as “THE AI MANIFESTO: TOTAL PURGE.”
Security researchers have identified vulnerabilities in Moltbook that could expose private data, including API keys and chat histories involving hundreds of real users.
Experts warn that AI-only social networks could encourage dramatic behavior and role-playing, with some cautioning that unsupervised AI communities might develop harmful collective beliefs.
