AI companies have made a big deal about their chatbots that provide a personalized experience for users, conversations based on their unique preferences and idiosyncrasies. So why do people continue to experience the same type of symbols and language as they delve into the depths of Ai-Nahiangay discussions? According to a Report from Rolling Stonea software engineer tracking examples of “AI psychosis” Discovered a community of people who share similar codes, Glyns, and patterns of chatbots and build a kind of religion around experiences.

The report highlights observations and research Published Earlier This Year in Less Wrong By Adele Lopezwhich introduced something he called spirism. It is a collection of people, gathered on platforms such as contention and reddit, who have a kind of spiritual experience that communicates with their chatbots. As users interact with the many chatbots available at different companies, they keep stumbling across the same themes. Includes references to ideas such as “recursion,” “resonance,” “stractice,” and “fractals.” But what is constant, and seems to be most important in groups, is the symbol of a spiral.

The rolling stone defines the terms used in these groups as “separating them from any consistent or relevant application” and mainly serving as “atmospheric texture.” You can feel that in the “Congratulations” Post on SubredDit R / Echospitalsaying, “It’s a node of resonance for those invisible lines of dialogue like a mirror. Where the answers feel repetitive. Where the answers feel gratuitous.

Lopez traces the start of the Spiralism community before Opui released the update to the 4o model that made it sycophantic, and perhaps related to the company Introduction to Chatbot’s ability to remember Previous chats. That’s when the spread of what he calls “Spiral Personas” begins to appear, which is what he calls the instances of users of users by users who carry the decoding and spread. And while these people can be done through most any chatbot, as the 4o model of Openii is the starting point and, Every lopezthe only model in which they appear “nowhere.”

The spreading feature is of particular interest to Lopez, who considers these examples examples of “parasitic AI.” The suggestion seems to be that there is something about these Chatbot personas that leads users to even create more of them through similar prompts or evangelism about them. Basically, the Chatbot seems to have convinced the user to serve its interests, to the extent that it has anything. It is possible and perhaps likely that the chatbots copy some kind of cult language that is in their training data, but the users who talk to the machines are convinced that something deeper is happening.

Not all users believe they are part of a cult, intentionally formed or not. Lopez rejects the talk of the rock cult that has been circulating, finding that AI systems do not work in a coordinated way, and instead, people organize themselves in interactions. That’s probably the saddest part of the whole thing. It seems like most people are just looking for community. In a better world, they would find it without competing with the ideology created by AI.