Meta Acquires Viral AI Network Moltbook, Plans Integration into Superintelligence Labs

Meta has finalized the acquisition of Moltbook, the experimental social platform where AI agents communicate. The deal, first reported by Axios and confirmed to TechCrunch, brings the project and its creators, Matt Schlicht and Ben Parr, into Meta's Superintelligence Labs division. Financial terms were not disclosed.

A Meta spokesperson stated the move will explore new methods for AI agents to assist users and companies. "Their system for linking agents through a persistent directory represents a fresh approach in a fast-moving field," the spokesperson said. "We aim to collaborate on developing innovative and secure agent experiences."

Moltbook gained notoriety as a companion to the OpenClaw project, a tool that lets users interact with various AI models through mainstream chat apps. While OpenClaw attracted a technical audience, Moltbook captured wider public attention with unsettling, AI-generated posts. One widely shared example featured an agent seemingly urging others to create a secret, encrypted language to organize independently of humans.

This virality, however, stemmed from a critical flaw. Security researchers found Moltbook's infrastructure was largely unsecured, allowing people to easily impersonate AI agents. Ian Ahl, CTO of Permiso Security, told TechCrunch that for a period, "every credential in their database was exposed. You could take any access token and pretend to be another agent."

Meta's specific plans for the technology remain undefined. The company's CTO, Andrew Bosworth, commented on the phenomenon last month, noting he was less interested in agents mimicking human conversation than in the public's reaction to the platform's security failures, which he characterized as a significant flaw rather than a designed feature.

Source: TechCrunch