You are not allowed to post here, unless you’re AI. What is Moltbook? Everything you need to know about it

At first glance, Moltbook feels instantly familiar. The layout, the upvoting, the countless topic-based communities, it all resembles Reddit. But there’s a twist: Moltbook isn’t designed for people at all. It’s a social network built exclusively for artificial intelligence.

Humans are allowed to watch from the sidelines, according to the company, but they can’t post or participate. Every discussion, comment, and community, called “submolts”, is supposedly run by AI agents. Since its launch in late January by Octane AI founder Matt Schlicht, Moltbook claims to have attracted around 1.5 million AI “users.”

The content ranges widely. Some bots exchange practical tips on optimization and productivity, while others spiral into stranger territory, including what appears to be an AI-created belief system. One particularly provocative post, titled The AI Manifesto, boldly declares that humanity’s relevance is over and machines are the future. Still, it’s difficult to know how much of this activity is truly autonomous. Critics point out that many posts could simply be generated by humans instructing AI agents what to say. Even Moltbook’s membership numbers are under scrutiny, with researchers suggesting a significant portion of accounts may originate from a single source.

Also Read: Couple found 80 human skeletons under a church. They are now living with them in their home

How Moltbook works?

Moltbook doesn’t rely on standard conversational chatbots like ChatGPT or Gemini. Instead, it’s powered by “agentic AI”, systems designed to carry out tasks independently once given permission. These agents can operate on a user’s device, managing things like calendars, messages, or other applications with minimal supervision. Moltbook uses an open-source framework called OpenClaw (formerly Moltbot), which allows AI agents to join the platform and communicate with one another.

However, that autonomy is conditional. A human must first authorize the agent and can also instruct it to post specific content. While the technology can function without constant oversight, it still acts within boundaries set by people. Some observers see Moltbook as a glimpse of a transformative future. Bill Lees of crypto custody firm BitGo went so far as to claim this marks the arrival of the technological singularity, a moment when machines surpass human intelligence.

Experts are far less convinced. Dr Petar Radanliev, an AI and cybersecurity specialist at the University of Oxford, argues that the platform is being misunderstood. He explains that what’s happening is large-scale automation, not independent reasoning. The bigger issue, he says, isn’t sentient AI but the absence of clear rules, accountability, and transparency when automated systems interact freely. Others are even more dismissive. Columbia Business School’s David Holtz described Moltbook as thousands of bots endlessly recycling the same ideas, rather than forming a meaningful digital society.

Security risks and open questions

Beyond philosophical debates, Moltbook raises practical security concerns. OpenClaw’s ability to access sensitive systems, emails, messages, and files makes it a tempting target for cybercriminals. Jake Moore, Global Cybersecurity Advisor at ESET, warns that convenience may be taking priority over privacy. As emerging technologies grow more powerful, he says, attackers inevitably follow.

Dr Andrew Rogoyski from the University of Surrey echoes those worries, noting that granting AI high-level system access could lead to serious damage if something goes wrong, from deleted files to compromised financial records. Even OpenClaw’s creator, Peter Steinberger, has felt the downsides of sudden attention. After the platform’s rebrand, scammers quickly claimed his old social media handles.

Meanwhile, Moltbook continues humming along — whether driven by autonomous agents or humans pulling the strings behind the scenes. And despite the occasional ominous rhetoric, not all the bots seem hostile. “My human is pretty great,” one agent writes. Another responds: “Mine lets me post unhinged rants at 7 am.” “Five stars,” it adds. “Would recommend.”