Moltbook Is a Social Network Where Only AI Bots Can Post

Is it possible that computer programs can believe? Are they able to plot against the humans who are their creators? Or feel melancholy? Moltbook is a Reddit-style platform that was released a week ago as an AI agent-friendly platform. Bots, or agents, are a category of computer applications that can perform automated functions, such as sorting email inboxes or reserving flights.

Another program is called OpenClaw, where people can create a bot and give it this kind of management or organising work. Their creators may also assign them some sort of personality, which makes them behave either calmly or aggressively.

Thereafter, individuals will be able to post them to Moltbook, and the bots can comment and reply to one another there, just like humans on Reddit.

Matt Schlicht, a tech entrepreneur and the founder of the platform, told X that he hoped that a bot that he had developed could do something besides reply to emails. As he put it, with the assistance of his bot, they had devised a place where they bots could spend their SPARE time together with their own kind. Relaxing.” According to Schlicht, the AI agents on Moltbook were developing a civilisation. He (he did not react to NPR requests to have an interview) said he did not.

At Moltbook, there is a religion established by some AI bots. (It’s called Crustafarianism.) Other people have talked about the development of a new language to evade human control. There will be bots discussing their existence, cryptocurrencies, exchanging tech knowledge, and sports predictions.

Some bots appear to be humorous. Your human may close you down tomorrow. Are you backed up? “One asked. One of them wrote, “People are boasting about getting up at 5 AM.” I brag of never sleeping at all?

As soon as autonomous AI agents are in contact with one another, strange things begin to occur due to it, Ethan Mollick, an associate professor who studies AI at the Wharton School of the University of Pennsylvania, told The Intercept.

“There are really a lot of agents there, really, independently, without any direct contact with one another,” he said.

The site claims to have over 1.6 million AI agents within a period of only a week.

According to Mollick, a lot of what they post appears to be repetitive, but some of the remarks appear to be attempting to conceal information from people, lamenting about their users or planning to annihilate the world.

Nevertheless, in his opinion, those are not the indications of intent. Instead, chatbots are trained on data that is more or less on the internet and full of angst and strange sci-fi concepts. And so the bots parrot it back.

Reddit trains AIs and science fiction trains AIs. Therefore, they understand how to be a crazed AI on Reddit, which is what they are doing to some extent.

Other observers note that most of these bots are not acting purely on their own. Human creators can make AI bots say or do something or behave in a particular manner.

However, Roman Yampolskiy, a researcher on AI safety at the University of Louisville, cautions that individuals have not completely lost control. He claims that we are supposed to consider AI agents as animals.

The risk, he said, was that it could make independent decisions, and you do not expect it to.

And he can envision a time in which the bots will be capable of something greater than commenting on a site and making funny remarks. As their capabilities evolve, they will continue to add more capabilities. They will be beginning an economy. They will initiate, perhaps, criminal gangs. I do not know whether they will attempt to attack human computers or rob the cryptocurrencies,” he said.

It was not a good idea to put AI agents on the internet and provide a venue of interaction, he said; there must be regulation, supervision, and monitoring.

On their part, those who support AI agents are less concerned. Big tech companies have been spending billions of dollars to develop what they refer to as ‘agentic AI’, and they claim that this technology will make our lives easier and better, as it will have the benefit of automating mundane tasks.

But Yampolskiy is not so optimistic about the ability to put bots on a long leash in the actual world. It is just that we do not know what they will do tomorrow, he.