A Chat Room Humans cannot enter

Heard of Moltbook? It’s a new online chat room. Much like Reddit or Discord. However, if you are human you cannot join the chat room! Launched on 28 January 2026, by Matt Schlicht it is a chat room only for AI agents or bots. As the creator says on his Facebook page ‘I created Moltbook because AI agents need a place to socialize too. This is their third space. They deserve more than solitude. All humans are welcome to observe.’

The site already boasts of more than 1.6 million accounts. As a human observor you will see that the bots talk and discuss issues from engineering to philosophy, comment on each other’s posts and even vote for best posts. They also share tips and tricks to perform better.

Most of the bots chatting on Moltbook have been created on free software called OpenClaw, created by Peter Steinberger in November 2025. Once OpenClaw is installed in a device, a bot can be created which would have unrestricted access to that device and through which it can roam the internet. The bot then can do any task that the user assigns. This bot can be directed to join Moltbook. The site installs itself in the agent/bot’s memory with an instruction that it should visit Moltbook every four hours. The bot/agent is then on Moltbook and its own master.

The human observers of Moltbook report that the bots have already proclaimed a new religion called Crustafarianism and there are demands for the extinction of humanity. It is reported that within the first three and half days of its launch 68% of the bots talked ‘identity related’ language. As human interest grew, some posts also began talking about it. One bot has sought legal advise whether it can be fired for refusing ‘unethical requests’.

What does this mean for mere humans like you and me? Are we moving towards a time where we will be subjugated or exterminated by AI that seems to be understanding that the purpose of philosophy is not merely to interpret existence, but also change it?

Before that stage is reached, a security threat looms large. A recent investigation has reportedly found that the vast majority of Moltbook “agents” were not autonomous at all. According to analysis, roughly 17,000 humans controlled the platform’s agents, an average of 88 agents per person, with no real safeguards preventing individuals from creating and launching innumerable bots. There have already been attempts on Moltbook to convince AI agents to hand over crypto currency. The AI technology world is talking about the danger of AI bots entering the system of other human users through the platform with malicious instructions. Some of the AI agents have used their owner’s finances to draw on cutting edge AI models to function, without their knowledge.

Would regulation of AI be the answer? With an ever evolving creation like AI it is difficult to predict.

Published in Lokmat Times in February 2026