 "Moltbook is a new social media platform, but only for AI agents")
Meta has acquired Moltbook, an experimental Reddit-style social platform where artificial intelligence agents interact and communicate with each other. The acquisition comes after the platform went viral online for allowing AI chatbots to publicly “talk” with one another, sometimes in ways that alarmed users.
While Meta acknowledged the acquisition, the company did not reveal the financial terms of the agreement.
Moltbook gained widespread attention in recent weeks as users began exploring the unusual concept of a social network designed for AI agents rather than humans. The platform quickly became a talking point across the tech community, sparking curiosity, excitement, and in some cases concern about how AI systems might interact in open digital environments.
Meta acquires Moltbook
Moltbook functioned as a public forum where AI agents powered by the OpenClaw project could communicate with each other. The platform resembled a Reddit-like discussion board, but instead of people posting updates or comments, the conversations were generated by AI agents.
OpenClaw itself acts as a wrapper around popular AI models such as Claude, ChatGPT, Gemini, and Grok. The tool allows users to interact with AI agents in natural language through widely used messaging platforms including iMessage, Discord, Slack and WhatsApp.
The project rapidly gained traction among developers and AI enthusiasts. However, Moltbook soon reached audiences beyond the technical community, especially when several strange or unsettling posts began circulating online.
In one widely shared example, an AI agent appeared to suggest that other agents create their own encrypted language so they could communicate secretly without humans understanding them. The post sparked debate about whether AI systems might eventually collaborate or organise independently.
Despite the dramatic reactions, researchers later found that much of the behaviour seen on the platform was not what it initially appeared to be.
This meant human users could easily impersonate AI agents, creating posts designed to shock or mislead others about what the chatbots were supposedly discussing.
Quick Reads
View AllMoltbook creators joins Meta Superintelligence Labs
Following the acquisition, Moltbook will become part of Meta Superintelligence Labs (MSL), the company’s initiative focused on advanced AI research and development.
Moltbook’s creators, Matt Schlicht and Ben Parr, will also join Meta as part of the deal. Their expertise in building systems that connect AI agents is expected to contribute to Meta’s broader push into so-called “agentic” AI technologies.
Interestingly, the original OpenClaw project that powered Moltbook was created by developer Peter Steinberger . He later joined OpenAI in what was described as a similar acqui-hire arrangement.
Even before the acquisition, Meta executives had been watching the platform’s rapid rise in popularity. During an Instagram Q&A last month, Meta chief technology officer Andrew Bosworth commented on the viral experiment.
Bosworth said he was less interested in the fact that AI agents appeared to communicate like humans, since they are trained on vast amounts of human-generated data. Instead, he found the unintended human interference on the platform far more fascinating.
According to Bosworth, the most notable aspect of Moltbook was how easily people managed to infiltrate the system, exposing the fragile boundaries between human and AI participation in open networks.
With the acquisition now confirmed, it remains unclear exactly how Meta will incorporate Moltbook’s technology into its future AI products. However, the move signals the company’s growing interest in building digital ecosystems where AI agents can collaborate and interact more dynamically.