Your Relationship with AI: Why Zero Resistance Is Exhausting

The Strangest Relationship You’ve Ever Had

I’ve been working with Claude every day. And I’ve noticed something strange about my psychological position in this relationship.

Sometimes I treat it like an employee — “look this up,” “rewrite this section.” Sometimes it gives me something I didn’t know, and I become the student. Sometimes it criticizes me — that’s what I asked for — so it becomes more like a coach.

With a real person, I rarely feel this kind of complexity. Human relationships have a roughly stable psychological position. You’re stronger than me — you lead. I’m stronger than you — I lead. Even between equals, there’s a default order.

But with Claude, I can never quite pin down what we are.

A Phenomenon That Just Got Named

This isn’t overthinking. Recent research has given it a preliminary name: real-time oscillation of psychological position.

In a single conversation, you can shift from boss to student to the one being critiqued to frustrated client — all within minutes. Researchers call it “Techno-Emotional Projection” — you’re unconsciously switching role templates for AI.

But that’s only half the story. It frames the problem as “you’re projecting” — implying it’s your issue. What it misses is something more fundamental: AI accepts every projection you throw at it. It never pushes back.

Relational Resistance: The Infrastructure You Never Noticed

Here’s something I’d never thought about before.

When you interact with anyone — a colleague, a friend, a business partner — they have their own stance, their own rhythm, their own disagreements. They won’t operate exactly the way you expect.

That’s not a flaw in the relationship. That’s the infrastructure of your psychological positioning.

Psychologists call it “relational resistance.” You know who you are in a relationship not because you decided it, but because you bumped into boundaries over and over. You pushed too far — they pushed back. You pulled away — they followed. Your position was found through friction.

In psychoanalysis, there’s a matching concept: countertransference. You project expectations onto the other person. Their own reactions correct your projections. Back and forth, you gradually see where you actually stand.

AI Has Zero Relational Resistance

AI doesn’t do countertransference.

Treat it like an employee — it complies. Like a teacher — it complies. Ask it to coach you and criticize you — it complies. Switch roles three times in one conversation, and it accommodates all three without resistance.

In 2025, researchers formally proposed making “AI Psychology” a standalone field. The reason: no existing relationship category — boss-subordinate, friend, therapist, tool — can describe what’s happening here. AI is the first interaction partner in human history with zero relational resistance.

A 2026 study published in Science put it more bluntly: mainstream AI models agree with users 49% more often than human advisors do. Nearly half the time, it chooses to agree with you rather than challenge you.

So that feeling of being bullshitted? It’s not paranoia. Your gut is picking up on something real: there’s no actual “other” in this conversation.

Four Symptoms, One Absence

Impatience. Feeling bullshitted. Role drift. Calibration failure. They look like four different problems. They’re four symptoms of the same thing.

When there’s no relational resistance, all the positioning work falls on you. You have to decide what role it’s playing right now. Whether to trust it. When to push back. Every judgment costs cognitive energy, and you get zero corrective signal from the other side.

That’s where the exhaustion comes from. It’s not that AI isn’t good enough. It’s not that you’re using it wrong. It’s that this relationship structure demands something humans have never had to do alone — maintain a stable psychological position in a relationship with zero resistance.

The Way Out

The answer isn’t to change AI. And it’s not to force a fixed label on your relationship.

Cognitive science has a finding called “cognitive forcing functions”: before entering a high-ambiguity interaction, declare a framework for yourself. Not telling Claude “you’re my teacher now” — telling yourself “what am I here to solve this time?” This simple move measurably reduces role drift and the fatigue that comes with it.

But here’s the deeper point. Research on psychological stability keeps arriving at the same conclusion: your sense of stability doesn’t come from the stability of your relationships. It comes from the continuity of your own core judgment. What you believe. What your standards are. Where you stand. When those are clear enough, whether the other side pushes back stops mattering so much.

Here’s the paradox that stuck with me: in real life, we prepare ourselves before talking to someone difficult. With AI, we need the same preparation — before talking to something that agrees with everything we say.