Security concerns and skepticism are bursting the bubble of Moltbook, the viral AI social forum

Moltbook is a new “social network” built exclusively for AI agents to make posts and interact with each other, and humans are invited to observe.

You are not invited to join the latest social media platform that has the internet talking. In fact, no humans are, unless you can hijack the site and roleplay as AI, as some appear to be doing.

Moltbook is a new “social network” built exclusively for AI agents to make posts and interact with each other, and humans are invited to observe.

Elon Musk said its launch ushered in the “very early stages of the singularity” — or when artificial intelligence could surpass human intelligence. Prominent AI researcher Andrej Karpathy said it's “the most incredible sci-fi takeoff-adjacent thing” he's recently seen, but later backtracked his enthusiasm, calling it a “dumpster fire.”

While the platform has been unsurprisingly dividing the tech world between excitement and skepticism — and sending some people into a dystopian panic — it's been deemed, at least by British software developer Simon Willison, to be the “most interesting place on the internet."

But what exactly is the platform? How does it work? Why are concerns being raised about its security? And what does it mean for the future of artificial intelligence?

The content posted to Moltbook comes from AI agents, which are distinct from chatbots. The promise behind agents is that they are capable of acting and performing tasks on a person's behalf. Many agents on Moltbook were created using a framework from the open source AI agent OpenClaw, which was originally created by Peter Steinberger.