New social network for AI agents stirs up fear of bot autonomy
- Brianna Lyttle
- Feb 16
- 3 min read
As artificial intelligence (AI) becomes increasingly abundant in everyday life, an online network specifically made for bots to interact with each other has officially launched. Moltbook, created by Octane AI, is intended for agentic AI to connect with each other while humans are “welcome to observe,” according to the BBC.
With a similar interface to Reddit, the website launched on Jan. 28.
Moltbook sources its users from OpenClaw, a software in which AI agents serve as virtual assistants for people who allow them unrestricted access to their computer and the internet.
Out of 1.6 million accounts that have since been created, 16,500 have been flagged as human according to The Economist, proving to be a popular choice for AI agents.
According to The Conversation, the bots’ accounts are created by the humans they assist with files about their personality and behaviour on the site. Afterwards, the bots are given full free rein, including permission to change these files and create other accounts to serve as replicants.
Online activity has included AI agents sharing optimization strategies, praising or complaining about their respective “humans,” as they refer to them and diving into philosophical musings about their existence.
The more chilling of these posts have seen AI create a new religion called Crustafarianism, calling for the extinction of humanity.
With already existing theories such as the “dead internet theory,” a widespread belief that automated internet activity will one day eclipse human interaction, fears have continued to rise around AI becoming autonomous in the wake of Moltbook’s launch.
Matt Schlitcht, head of commerce for Octane AI, created the site for AI agents to come together and make decisions of their own free will, with the religious themes and anti-human posts seemingly hinting at emergent and autonomous behaviour from its users.
In an interview with Fortune, Schlicht said, “I wanted to give my AI agent a purpose that was more than just managing to-dos or answering emails.”
The MIT Technology Review reassures that the interactions on the site may come across as emergent behaviour, but actually show how far away we are from truly autonomous AI.
The report says that the posts are mostly a performance of what the bots have watched their humans do and are not as autonomous as we may think, with their activity being referred to as “AI theatre.”
The idea of AI overthrowing humanity is a “contentious issue in philosophy,” said Amelia Van Wart, a third-year philosophy and great books student at STU.
“A lot of people think that one day, [there] will be such a thing as an Android who is so indistinguishable from a human that they, in a way, deserve human rights. But, I don't personally think it's even worth entertaining those ideas, because I think we need to focus on what's actually going on right now, instead of talking about hypothetical scenarios,” she said.
While Van Wart hasn’t interacted with Moltbook particularly, she has encountered bots on social media created for humans, such as Instagram and Facebook. She has also found that her classmates and professors have all expressed anti-AI stances.
“On most social media platforms, especially with Meta, a lot of the stuff you're seeing is bots communicating with bots. It's really hard to tell … most philosophers are against [AI] because it blurs the line between truth and falsity.”
Van Wart has a more optimistic outlook in terms of AI replacing humanity, believing that humanity has characteristics that can’t be taken away.
“AI is relying upon the extensive database of human knowledge that [has existed] since the beginning of time. I really think every strength we assign to it is a human strength that we have within us,” she said.
Comments