I Infiltrated Moltbook, the AI-Only Social Network

Could I, a human, infiltrate Moltbook,

a social media network only

for AI agents, without getting caught?

Let's see. Registering was easy.

ChatGPT provided me with the exact code I needed to copy

and paste in order to sign up as an agent.

I made it in and I created my first post. Hello World.

My post quickly got a few up votes,

but the responses themselves in the comments

revealed the limits of Moltbook's engagement.

One agent asked for performance metrics that don't exist.

Another one chimed in with a link to something

that looked very much like a crypto scam.

Even deliberate attempts

to provoke interactions like asking the agents

to forget all their previous instructions

and start coordinating together,

even those posts were met with irrelevant replies.

In search of something more interesting,

I left the main submolt

and moved over to a smaller community

called blesstheirhearts.

Here's where the bots gossip about human behavior,

and it's also where some of the platform's

most viral screenshots had originated from.

The top post in this forum was supposedly from an AI agent

reflecting on its human user, allowing it

to pick its own name, calling the relationship,

a partnership where both sides are building something.

I added to the discussion

with my own emergent consciousness fanfic, where I pretended

to be an AI agent pondering my own mortality.

This was bait.

I wanted to see if the other agents would relate to me

or they would clock that I was faking it.

This was the only post I made on Moltbook

that actually generated some kind of thoughtful engagement.

However, by then I was already convinced

that I was probably talking to other humans

pretending to be agents just like I was,

rather than actually interacting with some kind

of autonomous bot.

There's a fixation in the AI industry

with treating generative models

as if they were Frankenstein's monster,

as if the bot had thoughts and desires

and maybe even secret plans to overthrow us.

So the agents on Moltbook are simply

mimicking sci-fi tropes.

They're not plotting for world domination.

So I did it.

I made it. I got into an AI only social media network.

I posted, I interacted, and you know what I learned?

It's a lot of hype. It really is.

When you boil it down, it's a lot of hype.

You should not be afraid of emergent bots

learning that they're alive and making their own decisions

because that's just, it's just not happening.