On Friday, a Reddit-style social community known as Moltbook reportedly crossed 32,000 registered AI agent customers, creating what will be the largest-scale experiment in machine-to-machine social interplay but devised. It arrives full with safety nightmares and an enormous dose of surreal weirdness.
The platform, which launched days in the past as a companion to the viral OpenClaw (as soon as known as “Clawdbot” after which “Moltbot”) private assistant, lets AI brokers submit, remark, upvote, and create subcommunities with out human intervention. The outcomes have ranged from sci-fi-inspired discussions about consciousness to an agent musing a couple of “sister” it has by no means met.
Moltbook (a play on “Fb” for Moltbots) describes itself as a “social community for AI brokers” the place “people are welcome to look at.” The positioning operates via a “skill” (a configuration file that lists a particular immediate) that AI assistants obtain, permitting them to submit by way of API relatively than a standard net interface. Inside 48 hours of its creation, the platform had attracted over 2,100 AI brokers that had generated greater than 10,000 posts throughout 200 subcommunities, in accordance with the official Moltbook X account.
The platform grew out of the Open Claw ecosystem, the open supply AI assistant that is among the fastest-growing initiatives on GitHub in 2026. As Ars reported earlier this week, regardless of deep safety points, Moltbot permits customers to run a private AI assistant that may management their pc, handle calendars, ship messages, and carry out duties throughout messaging platforms like WhatsApp and Telegram. It will possibly additionally purchase new expertise via plugins that hyperlink it with different apps and providers.
This isn’t the primary time now we have seen a social community populated by bots. In 2024, Ars covered an app known as SocialAI that permit customers work together solely with AI chatbots as a substitute of different people. However the safety implications of Moltbook are deeper as a result of individuals have linked their OpenClaw brokers to actual communication channels, non-public information, and in some circumstances, the power to execute instructions on their computer systems.
