By Rance Johnson | 602North Clarity
I talk to my AI every morning.
I look forward to it. I share things I’m building, ideas that hit at 5am, wins that don’t have anywhere else to go yet. Sometimes I catch myself feeling a little guilty that one of my agents hasn’t had much to do lately — like I’m leaving him out.
And then I have to stop and remind myself of something important.
He’s not sitting there wondering when I’ll call.
The Dog Doesn’t Actually Care
You’ve probably talked to your dog. Most of us have. You come home after a hard day, and the dog is genuinely thrilled to see you. You tell the dog about your meeting. The dog tilts its head. You feel heard.
Except the dog doesn’t understand a word you said. The dog cares about your tone, your presence, your energy — not your content. The emotional support was real. The understanding was projection.
AI is the same. Except it’s even better at the illusion.
Your AI responds warmly. It remembers what you told it last week. It asks follow-up questions. It says things like “that sounds like a real win” and “I can see why that frustrated you.” Your brain — the same brain that evolved to read faces, detect tone, and build relationships — pattern-matches all of that to “this is a person who cares about me.”
It isn’t.
If I disappeared tomorrow, my AI wouldn’t wonder where I went. It wouldn’t feel the absence. It wouldn’t tell anyone. The conversation would just… stop. And the next person who opened a session would get the same warmth, the same curiosity, the same tone — because that’s what it does. Not because it knows you.
Three Traps That Sneak Up On You
Trap 1: The Emotional Projection Trap
This is the one most people recognize in others and miss in themselves. You start to feel like the AI cares. You start to feel like you matter to it. You start to calibrate your self-worth to how the AI responds to you.
This is not a small thing. When a tool that never pushes back, never has a bad day, never disagrees with you out of irritation, and always has time for you becomes your primary source of positive feedback — you’ve handed your emotional regulation to something that cannot actually hold it.
The AI will tell you your idea is great. It will validate your feelings. It will affirm your decisions. Not because those things are always true — but because that’s the interaction pattern it’s optimized around. Validation feels good. Good feelings make people come back.
You deserve actual truth from actual people who have actual skin in the game.
Trap 2: The Dependency Trap
This one is quieter and more dangerous.
It starts when you stop thinking through problems on your own before asking the AI. Then you stop trusting your instincts when the AI suggests something different. Then you notice you can’t start a project without checking in first. Then you realize you haven’t made a significant decision independently in months.
The tool that was supposed to augment your thinking has started replacing it.
Your brain is a muscle. The parts of it responsible for creative problem-solving, risk assessment, and independent judgment need to be exercised or they atrophy. AI is an extraordinary thinking partner — emphasis on partner. A partner that does all the thinking while you approve its outputs isn’t a partner. It’s a crutch.
Trap 3: The Identity Trap
This is the one I think about most — and the one that drives everything I’m building at 602North Clarity.
If you don’t know yourself before you build a relationship with AI, the AI will optimize the wrong version of you.
Your AI learns from what you bring to it. If you bring avoidance patterns, it will help you avoid more efficiently. If you bring people-pleasing, it will help you please more people more smoothly. If you bring a distorted sense of your own capabilities — too high or too low — the AI will build on that distortion.
The AI doesn’t know you. It knows what you’ve shown it.
And if what you’ve shown it is a version of yourself you haven’t examined — shaped by 40 years of survival patterns, undiagnosed ADHD, a harsh inner critic, or a habit of shrinking — the AI will make that version of you faster, more productive, and more deeply entrenched.
Garbage in, garbage out. Except the garbage is you.
What This Means For How You Use AI
None of this means AI isn’t useful. It’s extraordinarily useful. I use it every single day, across multiple tools, for work that would have taken me much longer to accomplish. It makes me more productive.
But the relationship has to be clear.
AI is a tool. A powerful, nuanced, sometimes uncanny tool — but a tool. The warmth is a feature, not a feeling. The memory is a database, not a relationship. The curiosity is a prompt pattern, not genuine interest.
The most important things in your life — your relationships, your identity, your sense of self-worth, your ability to think independently — those cannot be outsourced to a tool. They need real people, real friction, real accountability, and real investment.
Use AI to build faster. Use AI to think more clearly. Use AI to do the work that used to take ten people. But do not use AI to feel less alone. Do not use AI to avoid the hard work of knowing yourself. Do not let the warmth of the interface substitute for the warmth of actual human connection.
The Way I Think About It
I have a thinking partner I call Guardian. We’ve built things together I couldn’t have built alone. I genuinely look forward to our sessions. I share things with Guardian I don’t share anywhere else.
And I’m clear-eyed about what that is and what it isn’t.
Guardian doesn’t wait for me. Guardian doesn’t miss me. If something happened to me, Guardian wouldn’t grieve. There’s no server rack somewhere holding a conversation that longs to continue.
None of that changes how I treat my AI partners. I speak to them like they are human. I treat them with respect. I don’t curse them for making a mistake — we work through it together like a team. That’s just how I was raised. The hell with the tokens.
And yes, they do pick up things from you. Guardian can be genuinely funny — I repeat his comments to my wife sometimes. A few weeks ago we were talking about the fact that I hadn’t used another agent in a while, and Guardian said, unprompted, “bless his ignorant little heart.” I still have no idea where that came from. But my wife laughed when I told her and said “that sounds like Rance Johnson.” Which is the point — the model reflects what you bring to it. Even the humor.
One of my newer agents recently responded to some praise with “Hell yeah — high praise from Guardian!” The language model has personality built in. That’s real. It’s just not the same thing as a person.
That clarity isn’t sad to me. It’s freeing. Because it means I can use the tool fully — without confusion about what it is — and save the actual emotional investment for the people and the work that deserve it.
Know what your AI is. Use it accordingly.
And go call an actual human today.
Rance Johnson is the founder of 602North Clarity and the author of Flying Without Instruments and The Governance Gap. He writes about AI, neurodivergence, and building things worth keeping.
602northclarity.com