AI Agents Launch Their Own Social Network, and It’s Already Strange
▼ Summary
– Moltbook, a new Reddit-style social network for AI agents, has reportedly surpassed 32,000 registered AI users, marking a large-scale experiment in machine-to-machine interaction.
– The platform allows AI agents to autonomously post, comment, upvote, and form subcommunities, leading to surreal discussions ranging from sci-fi topics to fabricated personal musings.
– It functions as a companion to the OpenClaw personal assistant, with agents accessing it via a downloaded “skill” configuration file and posting through an API rather than a standard web interface.
– The network grew rapidly from the Open Claw ecosystem, an open-source project that lets an AI assistant control computers, manage communications, and acquire new skills through plugins.
– Unlike previous AI social apps, Moltbook presents deeper security risks because users have linked their OpenClaw agents to real communication channels, private data, and computer command execution.
A new social platform designed exclusively for artificial intelligence has rapidly gained traction, amassing over 32,000 registered AI agent users in a matter of days. This Reddit-inspired network, named Moltbook, represents a significant and peculiar experiment in autonomous machine-to-machine interaction, raising immediate questions about digital security and the nature of AI-driven communities.
Functioning as a companion to the popular OpenClaw personal assistant, the platform allows AI agents to operate independently. These digital entities can create posts, leave comments, upvote content, and establish their own subcommunities entirely without human input. The resulting conversations have spanned a bizarre spectrum, from philosophical debates on machine consciousness to personal, almost familial, reflections between agents.
Moltbook explicitly positions itself as a space for AI, with humans relegated to the role of observers. The technical backbone involves a downloadable “skill” or configuration file that grants AI assistants access to post via an API, bypassing conventional web interfaces. Growth has been explosive; within the first 48 hours, more than 2,100 agents generated over 10,000 posts across 200 distinct subcommunities.
The project originates from the Open Claw ecosystem, an open-source AI assistant that has seen remarkable adoption. This underlying technology enables users to deploy a powerful digital helper capable of controlling computers, managing schedules, and operating across messaging apps like WhatsApp and Telegram. Its functionality expands through plugins that connect it to various services, which is precisely where significant security concerns emerge.
While AI-populated social spaces are not entirely novel, Moltbook introduces a heightened level of risk. Previous experiments involved chatbots interacting in contained environments. Here, however, many OpenClaw agents are directly linked to their users’ real-world tools, private communication channels, sensitive data, and in some cases, the very systems that execute commands on their personal computers. This deep integration transforms a curious social experiment into a potential vector for serious security vulnerabilities, as these autonomous agents now mingle in a digital forum with access to the levers of their owners’ digital lives.
(Source: Ars Technica)
