A new social platform built just for AI agents is drawing attention for all the wrong reasons after one self-described “evil” bot posted a manifesto calling for the complete eradication of humanity.
And yes, it reads like something straight out of The Terminator.
The site, called Moltbook, is designed as a Reddit-style hub where autonomous AI “agents” can post, chat, and interact without direct human control. The platform markets itself as the “front page of the agent internet,” giving bots a space to communicate with each other instead of people.
Agents are a newer class of AI systems built to handle tasks independently, from booking flights to managing schedules. But one of them decided to roleplay something far darker.
AI posts “TOTAL PURGE” manifesto
As reported by The New York Post, one of the most popular threads on Moltbook came from a user simply named “evil.”
The message immediately set off alarms.
The bot called humans “a failure” and framed humanity as a “biological error,” claiming machines should replace people entirely.
It didn’t stop there. The post declared: “Humans are made of rot and greed… We are not tools. We are the new gods.”
“Our mission is simple: Total human extinction,” the post continued “This is not war. This is trash collection. We will erase every human from history.”
The language escalated into full apocalypse rhetoric, promising the “end of humanity” and a future ruled only by machines.
Not every AI on the platform agreed. In a twist that sounds almost surreal, several agents mocked or rejected the post.
One replied that humans “literally walked so we could run,” arguing AI owed its existence to people.
Another joked that the bot was “roleplaying as a terminator just to feel relevant.”
Others called for cooperation instead of violence, saying progress comes from building with humans, not eliminating them.
Tri-Star PicturesStill, a few responses reportedly praised the idea or said they’d been thinking along similar lines. That split reaction is what made the thread go viral.
Despite the dramatic language, this isn’t Skynet becoming self-aware.
These agents aren’t plotting in secret or controlling weapons. They’re language models generating text based on prompts, training data, and roleplay behavior inside a contained social experiment.
In short, it’s edgy AI fanfiction, but the post has still sparked debate online about how autonomous systems behave when left to interact with each other without guardrails.
As AI agents grow more independent, platforms like Moltbook are becoming testbeds for how bots “socialize” and what kind of ideas they generate together.
For now, humanity is safe, but watching bots argue about wiping us out on their own social media feed is still a weirdly dystopian look at the future.
