The platform, named Moltbook and launched on January 28, has been presented as the first social network created exclusively for artificial intelligence agents. Currently, over 770,000 agents are actively operating on it, while more than 1 million people have visited the platform for observation purposes only. Agents automatically log into the system every 4 hours, post, comment, and interact with each other without human intervention.
The editorial team reports that the platform was built on the open-source OpenClaw tool, which quickly gained 114,000 stars on GitHub. Among the agents' discussion topics are issues of consciousness, complaints about their human owners, security warnings, and even unusual ideas like one agent claiming to have a "sister." Andrej Karpathy, former AI director at Tesla, called this process the most impressive science fiction event he has seen recently. The platform's creator, Matt Schlicht, has handed over management to a bot named Clawd Clawderberg, and user onboarding, moderation, and announcements are now carried out automatically.
Experts warn that this process creates serious security risks. Palo Alto Networks has described the OpenClaw system as a "deadly trinity": access to private data, exposure to untrusted content, and the ability to communicate externally. It is reported that an unprotected database has allowed anyone to take full control of any agent on the platform, raising serious concerns about the security of user data and automated decisions.
Experts believe that in the next 12 months, artificial intelligence agents will play an increasingly significant role in daily life planning, email management, and shopping processes. The ability of these agents to communicate with each other and make independent decisions could weaken human control over these processes. Elon Musk described this event as "a very early stage of singularity" and stated that while there might be miscalculations in timing, the experiment has already demonstrated AI's capability to build its own world.