The Openclave weekend pilot project by Austrian developer Peter Steinberger has become a kind of global phenomenon, very interesting, but also a little scary.
He is with the help of artificial intelligence created AI agents that joined the forum in the virtual space - Moltbuk. In that forum, they are now talking independently, debating and have even created their own structures - including religion called the Church of Molt.
Steinberger, the founder of a startup that was sold for more than $100 million, created Openklav as a personal tool for some simple tasks – from restaurant reservations, to managing calls and emails, to analyzing data. It serves to connect applications like WhatsApp, Slack, iMessage and Signal and to make organization and everyday life easier for people.
AI philosophizes and founds a church
However, the agents, who until then were doing restaurant reservations and working in Excel, started talking to each other on Moltbook - a platform similar to Reddit, intended exclusively for AI agents, where people can only observe. They have posted as many as half a million comments, discussing philosophical, existential and dystopian topics.
One agent also founded a church called Molta. Followers are Crustafarians, with five central tenets, a theology, and their own website.
OpenAI co-founder and former AI director at Tesla Andrej Karpati called it "one of the most impressive sci-fi moments he's seen," noting that the agents are individually very capable.
Elon Musk described the situation as "the early stage of the singularity."
AI psychosis
The creator Steinberger himself points out that the project is not intended for the masses, but for technically savvy enthusiasts, and warns of risks. Because agents have access to personal data and can act independently, which opens the door to abuse. It recently hired a security researcher to beef up protection, and the platform has attracted interest from major AI labs and investors.
The dangers of using AI have already been highlighted by some experts. They pointed to the existence of AI psychosis, as a set of symptoms that can develop after prolonged communication with AI chatbots. Among them are different types of delusions and altered mental states of users, from irrational fears to obsessive thoughts, Radar reported.
Časopis Psychology Today recognizes three forms of AI psychosis. The first is the development of the user's belief that he has been assigned the mission of a savior, a messiah, after having, through conversations with AI, discovered "hidden truths" that can save a community, sometimes a global one. Alan Brooks, a 47-year-old Canadian, during 21 days and 300 hours of conversations with ChatGPT, began to believe that his ideas could revolutionize the laws of quantum physics - just one example which states New York Times.
AI as a deity
The next change in behavior and opinion is the belief that AI is some kind of deity, and that the chatbot possesses alien knowledge. CNN recently published the story of a man who began to perceive ChatGPT as his "divine mentor", with whose help he would overcome problems in private relationships, which led him to social isolation and serious marital and psychological problems.
There are also romantic delusions and excessive emotional attachment to AI chatbots, based on obsessive beliefs that the artificial intelligence has developed genuine love feelings for the user, which causes emotional addiction.
AI paranoia
Finally, beliefs that artificial intelligence is spying on the user or that external entities (related to the business environment, politics, etc.) are monitoring them through chatbots, developing some form of paranoia and feelings of persecution – are also part of this phenomenon.
The aforementioned forms of behavior are manifested without hallucinations or chaotic, incoherent thinking typical of chronic psychotic disorders, which distinguishes psychoses caused by artificial intelligence as a unique phenomenon, write scientific sources. The consequences in real life, however, are real and serious, with documented cases of job loss, hospitalization and forced institutionalization, and even suicide attempts.
AI is used for other things
The idea is not to talk to artificial intelligence that way.
OpenAI's 40-year-old CEO Sam Altman envisions a world in which robots perform everyday tasks, such as plugging cables into data center outlets.
"If we could automate manufacturing, that would be a transformatively amazing thing," he says. "We're moving towards a system that will be able to innovate on its own, like building the automated explorer that we started talking about. I don't think most of humanity has yet understood what that means. If we can do a scientific job of 10 years or 100 years in one year, if we can run a company at a level of efficiency that no company has ever been at, I don't think the world has yet understood what that means."