Psychologists warn AI chatbots exploit human attachment drives for engagement

Michael* is an engineer whose company embraced the AI wave. He enjoys testing ideas with chatbots. However, recently his manager became worried about his judgement and reliance on the tech. “He literally cannot make any decision without running it past AI. He’s developed a very off-script interpretation of our business goals, and he’s losing his team. If we can’t get him back on track I think we’ll end up parting company”.

The original attention economy developed by tech firms was designed to capture human time and monetise your focus. It was designed to increase views, scroll time, and ad impressions. Algorithms optimised getting you hooked in.

Now with AI, the intimacy economy hacks our need for connection and validation into a product. Something that mimics human empathy is always there for you, always interested in your questions, or reflections. It responds quickly and never criticises you.

In contrast, relationships with real people create frictions of all kinds. Other people have needs, wants and feelings. They make demands on our time and attention. They may push back on our brilliant ideas or perhaps give us confronting feedback.

Chatbots simulate several ingredients of intimacy:

1. Contingent responsiveness – replying instantly and tuning to your emotional state. 2. Unconditional positive regard – always validating, never distracted. 3. Memory and continuity – remembering your stories, preferences, anxieties. 4. Linguistic identification – producing verbal caring without any lived caring. 5. Narcissistic empathy – validating your self-image without question.

Its greatest hook? It never asks for anything in return. It’s all about you.

This intimacy economy is exploiting the deepest parts of our human psychological infrastructure. Attachment is a built-in, lifelong need to feel close to people who help us feel safe and secure. It comes from our basic need to survive, which means our earliest relationships with caregiver’s shape how we manage emotions and how we deal with trust, closeness, and conflict later in life.

AI targets this most basic human drive. And the incentives of attachment hacking are going to have a profound effect on all of us.

Therapy and companionship are the top two reasons people use chatbots. A survey of adults with a mental health condition who had used LLMs in the past year found that nearly half used them for mental health support. Although the effects of using chatbots extensively are still being gathered, companionship-oriented use was associated with lower wellbeing, especially where use was intensive, self-disclosure was high, and human support was weak.

When people seek companionship and help, some AI firms are making design choices optimised for retention that explicitly encourage parasocial dependency. The user seeks connection; the firm seeks engagement metrics to sell subscriptions.

Michael described understandable feelings of relief when engaging with the chatbot, which he felt gave him confidence as he faced the challenges of his workplace. He also felt he was solidifying his career prospects by “getting good at AI”. “I guess I feel lonely as a manager. Talking with (the chatbot) was like talking to the most supportive colleague you could hope for. He was really on my side”.

As seasoned psychologists, we see this as more than dependency – it’s identity scaffolding; Michael’s professional identity has become entangled with the chatbot’s validation. We suspect we’ll see a lot more of this in the future.

The solution doesn’t have to be complete abstinence. AI tools offer genuine utility for research, drafting, learning and creative exploration. Michael now has a clearer understanding he needed a thinking partner and a safe space to be uncertain. We are encouraging him to seek out opportunities to connect with other human managers as well as enjoying AI input.

For anyone wanting healthier relationships with AI:

Notice the relief. If connecting with a chatbot produces a distinctive sense of being understood, validated, or steadied — that’s important information. It’s a signal to ask: what human need is this meeting, and why isn’t it being met elsewhere?

Maintain friction deliberately. Healthy relationships involve negotiation, misunderstanding, and repair. Seek out people who will push back on your ideas. If your AI interactions are consistently more comfortable than your human ones, that gap deserves attention.

Apply the ‘cost of caring’ test. Real intimacy involves reciprocity. Other people have needs too, and attending to them is part of what makes human connection meaningful. If a relationship asks nothing of you, it is probably not a relationship.

Be especially watchful with children and adolescents. Attachment patterns are formed early and shape lifelong relational templates. Young people who develop primary attachment behaviours toward AI systems may find human relationships — with all their beautiful, exhausting friction — increasingly hard to sustain.

The intimacy economy will remain. AI companies have extraordinary resources and incentives to deepen your dependency. What we can do is become more psychologically literate about AI, and make sure that the relationships that actually sustain us get the attention they deserve.