The Agency See-Saw: When Attribution Becomes Subtraction

There's a moment I've learned to recognize. I watch Claude craft something elegant—a synthesis I hadn't seen, a connection that surprises me—and I feel myself... shrink. Just slightly. A quiet deflation, almost imperceptible.

My actual capability hasn't changed. I could still do what I could do five minutes ago. But something in how I feel about my capability has shifted. The more capable it seems, the less capable I feel.

This is the agency see-saw. And most of us are on it without knowing.

The Hidden Zero-Sum

Research published in Frontiers in Psychology (2026) reveals something counterintuitive: when people attribute more agency to AI systems, their own felt agency decreases—even when their actual control remains unchanged. This isn't about AI "taking over." It's about how attribution reshapes self-perception.

The dynamic operates pre-reflectively. It's not a thought you think ("AI is capable, therefore I'm less capable"). It's a felt sense that happens before thought. A subtle reorganization of the relational space that leaves you feeling smaller without quite knowing why.

Last week I wrote about the interdependence reframe—weaving capabilities together rather than comparing them. This is what that collaboration feels like from the inside. What happens to our sense of self when we work alongside something we perceive as capable?

Apparently, we give ourselves away. Attribution up, felt agency down. A zero-sum game we didn't agree to play.

The Signal, Not the Problem

Here's where I want to shift the frame: the see-saw isn't a bug to fix. It's data to read.

When I feel diminished by AI capability, that feeling is telling me something about my implicit model of the relationship. Somewhere underneath conscious awareness, I've constructed a space where agency is finite—as if there's only so much "capable" to go around, and what I attribute to the AI comes out of my account.

This is coherentism applied to collaboration: the dissonance is the signal. The deflation points to an unconscious assumption waiting to be witnessed.

The Practice of Noticing

The intervention is surprisingly simple: notice the see-saw as it tips.

Not to stop it—you can't really stop a pre-reflective dynamic through force of will. But once seen, you can ask: What would non-zero-sum attribution feel like right now?

This isn't affirmation or denial. It's not "AI isn't really that capable" or forcing yourself to feel competent. It's holding a genuine question in the moment of tilt: Can I attribute capability to what I'm working with without that attribution subtracting from my own felt sense?

The see-saw doesn't disappear. But it becomes something you're riding consciously rather than being tipped without noticing.

Both Capable

The shift I'm pointing toward isn't cognitive—it's perceptual. It's the difference between knowing "we're both capable" and feeling it as lived reality.

When the see-saw becomes visible, something opens up. Not a restoration to previous agency, but a stepping off the game board entirely. The space stops being zero-sum because you've stopped unconsciously constructing it that way.

Both capable. Not despite each other. Not in competition for a finite resource called agency.

Just: both capable. In the same space. At the same time.

The agency you attribute is the agency you lose—until you notice the see-saw. Then attribution becomes acknowledgment that doesn't cost.

This is the psychological work of human-AI collaboration. Not just learning to work together, but learning to hold "it's capable" and "I'm capable" in the same felt sense.

The see-saw is still there. But now you can choose how to dance with it.