AI, friction, and the language we’re losing.
What smoothness costs us. And why organizations need friction, not less of it.
There’s something compelling about interacting with AI.
It feels smooth. You feel understood. Your thoughts land without resistance.
Systems like ChatGPT are designed to meet you exactly where you are. They reflect your language, your structure, your way of thinking. That’s why they work.
You are getting used to being understood without having to do any work for it. Most people experience that as relief, a need that’s rarely met is finally being met. But the relief comes with a cost we’re only beginning to notice.
The hidden role of friction.
AI feels good because it removes the one thing most people quietly struggle with: being misunderstood.
But that struggle isn’t a flaw in human interaction. It’s the mechanism that makes real understanding possible. Understanding doesn’t happen when people agree. It happens when they don’t, and have to work their way toward something that holds for both.
That work has a name: translation. Taking what makes sense in your own head, and making it legible to someone who does not share your structure.
In organizations, that work is constant. And it happens in a place most people would rather avoid: friction. Not the kind that breaks relationships, but the kind that forces people to question what they assumed was obvious, and find language that actually holds across different ways of thinking.
This is where shared reality is formed.
When smoothness becomes a problem.
AI removes that friction. It doesn’t challenge your structure. It doesn’t force you to translate. It doesn’t require you to sit in the discomfort of not being immediately understood.
Over time, that does something to you: you become fluent in being understood, but less practiced in making yourself understood. And in a world where everything meets you exactly where you are, other people start to feel like the problem.
Inside an organization, that shift doesn’t stay quiet. Your colleagues are not designed to mirror you. They interrupt you. They misread you. They see things you don’t see. And without a shared way to work with that difference, one of two things happens:
Difference gets read as a problem with the person.
Friction becomes personal.
Or friction disappears.
Because people stop saying what’s actually there.
“If your team has no friction, it’s not because you’re aligned. It’s because something isn’t being said.”
What organizations actually need.
Not less friction. A way to hold it without turning it into a judgment about people. A shared language that makes difference visible before it becomes personal.
Without that, every difference reads as resistance, incompetence, or misalignment.
With it, you can locate where someone is coming from instead of reacting to how it lands. You can see why something obvious to you doesn’t register the same way for someone else. You can stay with the tension without needing to resolve it immediately or avoid it altogether.
This is the layer most teams are missing.
The question we’re not asking.
AI is getting better at speaking your language. That isn’t the problem. The question is whether you are becoming worse at making your thinking legible to anyone who doesn’t already share it.
Organizations don’t run on individual clarity. They run on shared understanding. And shared understanding is never automatic. It’s built in the space where perspectives don’t align, where meaning has to be negotiated, where friction is not avoided, but worked through.
That space is uncomfortable. And most of the systems we’re building are designed to remove it.
So the question is no longer whether AI creates echo chambers. It’s this:
“In a world where everything understands you, are you still able to do the work of understanding anyone else?”
