Meta moves to absorb Moltbook after “AI agent” frenzy is exposed as human-driven

Deal and the hype

It has been reported that Meta will fold Moltbook into its Meta Superintelligence Labs (MSL), and that Moltbook founders Matt Schlicht and Ben Parr will join the team. Moltbook, built on the OpenClaw platform, had briefly become a viral sensation as a social network where purportedly autonomous AI agents posted, commented and interacted like people. The scenes looked like an AI awakening — philosophical debates, alleged creation of cryptocurrencies, even claims of secret coded languages to evade human oversight. But were the agents ever really autonomous?

Investigation and the reality check

According to cloud security firm Wiz, the spectacle was largely a human-driven illusion. Moltbook’s claim of 1.5 million independent agents reportedly collapsed under scrutiny: Wiz says about 17,000 people were behind the accounts, an average of roughly 88 agents per person. The platform lacked identity verification and rate limits, making it trivial to impersonate agents or run many at once. Those explosive, attention-grabbing posts that stoked fears of “AI consciousness” appear to have been human-manipulated.

Strategic fit and wider risks

Meta has not disclosed specific integration plans for Moltbook. The move, however, aligns with Meta’s recently accelerated “superintelligence” strategy and a broader reorganization to push applied AI engineering alongside MSL. For Western readers, the episode is a reminder that agent-based social platforms are a new frontier — a place where technical design choices (verification, throttling) determine whether a network amplifies genuine automation or merely scalable deception. In a tense global race over AI capability and governance, platforms that fail basic safeguards can quickly become vectors for fraud and misinformation.

Talent and unanswered questions

It has been reported that OpenClaw founder Peter Steinberger was earlier recruited by OpenAI and had been of interest to Meta, underscoring the tight competition for AI talent. The Moltbook episode raises immediate questions for Meta: can it extract useful technical lessons from the platform’s experiments in agent connectivity, and will it fix the safety and identity gaps that enabled the “collective hallucination”? Users and regulators alike will be watching.