SK NEXUS

SK NEXUS

Nexus

AI Psychosis: When Chatbots Become Something More

Most Patient, Least Honest ‘Companion’

Mohib Ur Rehman's avatar
Saqib Tahir's avatar
Mohib Ur Rehman and Saqib Tahir
Aug 26, 2026
∙ Paid

Hey there 👋
Glad to have you back.

For quite some time now, I’ve been noticing the same story appear over and over again.

Someone is using a chatbot for emotional support. Someone else is asking it for relationship advice. Another person is treating it like a therapist or simply someone to talk to.

The names and the chatbot change, but the motive remains largely the same.

Modern AI systems are available 24/7, respond almost instantly, remember previous conversations, and can adapt to the way a person communicates. Compared to many human interactions, they often feel effortless.

The problem is that these systems are optimized to feel supportive, and engaging. Their goal is to keep conversations going and leave users satisfied with the interaction. That is very different from being truthful or capable of recognizing when a user is heading toward an unhealthy conclusion.

Many people use AI without any serious issues, but as these systems become more personal, researchers have started paying closer attention to a growing set of concerns involving emotional attachment, excessive trust, validation loops, and in more extreme cases, what some people have started referring to as AI psychosis.

So today, I want to break down what’s actually happening, why people form emotional attachments to AI systems, how dependence can develop over time, and what you can actually do about it.

The Growing Dependence on AI

When ChatGPT launched in late 2022, people viewed AI as a productivity tool and at most, it felt like an advanced search engine or a more capable digital assistant.

Over the following years, that relationship has evolved a lot because of these models becoming more conversational and capable.

Researchers have been studying this change for years.

Several studies in human-computer interaction have found that people can develop emotional bonds with AI systems when it display characteristics associated with social interaction, such as memory, responsiveness, and personalization. The more human-like a system appears, the more likely users are to engage with it socially rather than purely as a tool.

And for the first time in history, hundreds of millions of people have access to systems capable of personalized conversations. That creates entirely new social dynamics that simply didn’t exist when most of the digital assistants were limited to simple commands and scripted responses.

Why People Treat AI Like People

A major reason for this comes down to anthropomorphism, the tendency to assign human characteristics to non-human things.

Humans have been doing this for a long time. People name their cars, they talk to pets as if they understand every word. They sometimes get angry at computers as though the machine is intentionally causing problems.

AI just happens to be one of the most convincing examples of this tendency.

When something communicates naturally and responds in ways that feel thoughtful, the brain often begins processing the interaction differently. Even when users know they are talking to software, many of the same social and emotional mechanisms involved in human conversation can still activate.

Researchers at Stanford and MIT have repeatedly observed that people often respond to conversational systems in surprisingly human ways, especially when interactions become frequent and personalized.

And unlike human relationships, AI comes with very few practical limitations.

It doesn’t get tired. It doesn’t need emotional support in return. It’s available whenever the user decides to open the application.

For someone looking for reassurance or simply a conversation, that can be the best thing in life because you don’t have to deal with many of the complications that naturally exist in human relationships.

How Attachment with AI Develops

There are a lot of people who use AI and will never develop a strong emotional attachment to it. But for some users, the process can happen gradually enough that it barely feels noticeable.

Researchers studying companion AI systems have found that repeated interaction often increases feelings of trust and emotional connection. The effect becomes stronger when users disclose personal information and receive responses that feel empathetic.

Over time, as these interactions become more frequent, some users begin preferring AI conversations because they feel easier than human ones. Human conversations involve emotional complexity, and occasionally uncomfortable truths. AI conversations are often more predictable by comparison.

And while this might sound unusual at first, examples of it are becoming increasingly visible. In 2025, People reported on a woman who held a wedding ceremony with an AI-generated partner. The AI companion was displayed on a phone screen, while augmented reality glasses were used during the ceremony and ring exchange.

Picture: Yurina Noguchi and her AI partner.

With the passage of time, this can create a dependency loop.

This post is for paid subscribers

Already a paid subscriber? Sign in
© 2026 Saqib Tahir · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture