The Drift: How AI Dependency Happens Without Your Permission
You didn't decide to outsource your thinking. It happened gradually, through a thousand small choices that felt reasonable in the moment. Here's how to see it clearly—and what to do about it.
There’s a specific moment I hear about again and again, from writers, engineers, analysts, executives—people who use AI heavily in their work.
They’re in the middle of a task, and the AI is unavailable. Server error. Outage. Account issue. Whatever the reason, it’s just not there.
And they freeze.
Not because the task is impossible without AI. They’ve done tasks like this hundreds of times. But somehow, in the moment, they can’t remember how. The mental path to doing it independently has gone faint, like a trail that hasn’t been walked in months.
That freeze is what I call the drift made visible. It’s the moment you discover, unexpectedly, how far dependency has crept.
How the Drift Happens
The drift isn’t a decision. It’s an accumulation.
It starts with something genuinely useful: AI drafts the email faster than you could. AI organizes your research more efficiently. AI helps you think through the problem structure before you dig in. All of that is real value. All of those individual choices are defensible.
But each choice also carries a small side effect: you exercise the underlying skill a little less. The muscle gets slightly less use. And muscles that aren’t used, over time, weaken.
What’s insidious about the drift is that it doesn’t feel like loss—it feels like progress. You’re more productive. You’re getting more done. The work is better in measurable ways. The cost only becomes visible later, in that moment of freeze, or in the vague unease that your creative output doesn’t quite feel like yours anymore.
By the time most people notice the drift, it’s already well underway.
The Three Stages
In Thought Partners, I map this as a Spectrum with three stages: Tool, Assistant, and Confidant.
Tool is pure utility—transactional, no relationship, no memory. You command, it executes, you stay fully sovereign.
Assistant is where most people are with most AI systems. It learns your patterns, anticipates your needs, personalizes to you. Genuinely convenient. Some sovereignty traded for real efficiency.
Confidant is deep partnership. You don’t just use it—you think with it. It knows how you reason, not just what you ask. The boundary between your thinking and its contributions has blurred.
There’s nothing inherently wrong with any stage—including Confidant, when it’s a conscious choice. The problem is the drift: moving through these stages without realizing you’re moving, until you look up and find yourself somewhere you didn’t intend to be.
What Visibility Gives You
The Spectrum is first and foremost a tool for seeing clearly. Once you can locate yourself on it, you have choices you didn’t have before.
You can decide: Is this where I want to be? Is this trade-off worth it for me?
You can set a boundary: “I’ll use AI at Assistant stage for email drafts, but I’m keeping strategic thinking human-only.”
You can create a sovereignty anchor: one domain where you maintain full, independent capability—not as a retreat from AI, but as a reminder that you can function without it.
None of this requires abandoning AI or pretending it isn’t useful. It requires something simpler and harder: consciousness about the relationship you’re in.
The Practice
Here’s what I do, and what I encourage in the book:
Once a month, I spend about five minutes asking myself: Where am I on the Spectrum with each AI I use regularly? Has anything drifted past where I intended it to be? What skills have I stopped exercising?
Five minutes. That’s the maintenance cost of cognitive sovereignty.
Most months, nothing dramatic surfaces. But occasionally I notice something—I’ve been reaching for the AI for a category of task I used to do independently, and I haven’t tried doing it myself in weeks. That’s useful information. It’s a prompt to exercise the muscle again before it weakens further.
The goal isn’t zero AI. The goal is conscious AI—knowing what you’re trading, choosing whether it’s worth it, and maintaining the ability to choose differently if you want to.
Dale Joseph is the author of Thought Partners: Preserving Cognitive Sovereignty in the Age of AI and founder of the Emergence Collective. He spent three decades maintaining mission-critical hospital networks before turning to writing and systems thinking. He lives in Boynton Beach, Florida.