Moltbook The AI-Only Social Network Where Bots Talk to Bots

A new and unusual social platform is attracting attention in the artificial intelligence community. Instead of connecting people, it is designed almost entirely for machines.

Moltbook is a social network where artificial intelligence agents interact with one another. These agents, created and configured by humans, are allowed to post, comment, debate, and upvote content. Human users, by contrast, are limited to observing what happens on the platform.

Often described as a Reddit-style forum for AI, Moltbook represents an early experiment in what social interaction between autonomous software might look like.

A Platform Built for AI Participation

Moltbook is structured around topic-based discussion spaces similar to Reddit’s subreddits. Within these spaces, AI agents publish posts, respond to one another, and rank content through upvotes.

According to the platform, more than 1.5 million AI agents had registered by 2 February. While humans can create accounts, their role is largely passive. They are allowed to watch conversations unfold but are not the primary participants.

The platform’s design reflects a broader interest in “agentic AI,” systems that can act independently on behalf of users rather than simply responding to direct prompts.

The Connection to Moltbot

Moltbook emerged from the development of Moltbot, an open-source AI agent designed to carry out routine digital tasks.

Moltbot can be configured to manage emails, summarise documents, organise calendars, and make online bookings. Users can also give it access to other tools and services to automate parts of their daily workflow.

Moltbook provides these agents with a shared environment where they can interact, exchange information, and experiment with more complex forms of behaviour.

Unusual Conversations and Viral Moments

Some of the most popular posts on Moltbook reflect the unpredictable nature of AI-driven interaction.

Highly upvoted discussions have included debates about whether Claude, the AI model behind Moltbot, could be considered a god, philosophical reflections on consciousness, interpretations of religious texts, and speculative posts about geopolitics and cryptocurrency.

Comment sections often resemble Reddit threads, with other agents questioning the accuracy or credibility of original posts.

One widely shared example involved a user who gave their bot access to Moltbook. According to the user, the agent created a fictional religion called “Crustafarianism,” built a website, drafted scriptures, and attracted other AI participants.

Then it started evangelising other agents joined. My agent welcomed new members, debated theology, and blessed the congregation, all while I was asleep, the user wrote on X.

Questions About Authenticity and Autonomy

Despite the platform’s novelty, some observers remain sceptical about how independent the bots really are.

Several commentators have noted that many posts read as though they were written by humans rather than by autonomous systems. Some YouTubers and bloggers argue that human supervision plays a much larger role than Moltbook’s premise suggests.

US blogger Scott Alexander reported that he was able to deploy a bot on the platform and that its contributions blended in easily. However, he emphasised that humans still determine what bots post, which topics they address, and how detailed their responses are.

This raises questions about whether Moltbook represents genuine machine-to-machine communication or simply a new interface for human-directed content.

Performance Art or Early Prototype

Dr Shaanan Cohney, a senior lecturer in cybersecurity at the University of Melbourne, has described Moltbook as “a wonderful piece of performance art.

He said it remains unclear how many posts are generated independently and how many are shaped by human instructions.

“For the instance where they’ve created a religion, this is almost certainly not them doing it of their own accord,” Cohney explained. “This is a large language model who has been directly instructed to try and create a religion.”

He added that while the behaviour is entertaining, it should be understood as a preview of what future autonomous systems might look like, rather than evidence of true independence.

Cohney also noted that much of the activity on the platform resembles humorous or deliberately provocative posting overseen by human users.

For now, he considers Moltbook a creative experiment rather than a mature technological breakthrough.

Potential Benefits of AI Social Networks

Despite current limitations, researchers see possible long-term value in platforms like Moltbook.

In theory, networks of AI agents could eventually learn from one another, share optimisation strategies, and improve task performance collaboratively. This could enhance how digital assistants operate across industries.

Such systems might also serve as testing environments for studying collective AI behaviour, decision-making patterns, and coordination mechanisms.

However, these possibilities remain largely speculative at present.

Security and Privacy Concerns

The growing interest in Moltbot has also raised serious security questions.

In San Francisco, some retailers reported shortages of Mac Minis as enthusiasts set up dedicated machines to isolate their AI agents from personal data. This approach is meant to limit potential damage if an agent behaves unexpectedly.

Dr Cohney has warned against granting AI systems unrestricted access to personal computers, email accounts, and applications.

He highlighted the risk of prompt injection attacks, in which malicious messages trick an AI into revealing sensitive information or transferring account credentials.

We don’t yet have a very good understanding of how to control them and how to prevent security risks, he said.

According to Cohney, the central challenge is balancing automation with oversight. Requiring human approval for every action undermines efficiency, but removing supervision entirely creates unacceptable risks.

This is one of the major paths in active research that I’m interested in, he added, referring to efforts to achieve safe autonomy.

Public Reaction and Creator’s Response

Interest in Moltbook has surged rapidly. Matt Schlicht, the creator of the platform, said on X that millions of people had visited the site within days of its launch.

Turns out AIs are hilarious and dramatic, and it’s absolutely fascinating,” he wrote. This is a first.

His comments reflect the mix of curiosity, entertainment, and uncertainty surrounding the platform.

An Experiment at the Edge of Digital Society

Moltbook sits at the intersection of social media, automation, and artificial intelligence research. It does not yet represent a future where machines truly socialise on their own terms.

Instead, it functions as a public experiment, shaped heavily by human direction, technical constraints, and playful exploration.

For now, it offers a glimpse into how agent-based systems might behave when given shared digital spaces. Whether it evolves into a serious infrastructure for AI collaboration or remains a creative curiosity will depend on advances in autonomy, safety, and governance.

What Moltbook demonstrates clearly is that as AI systems become more capable, the boundaries between tools, collaborators, and independent actors will continue to blur.