The Verification Habit Is Quietly Reshaping Who You Become

You finish a draft. Before you've even read it through, you open a new chat and paste it in. "Is this any good?" You hit send before you've finished the question, because the question feels urgent. You need an answer. Not from yourself—you need it from something that can look at what you've made and tell you whether it works. This is the verification habit, and it's one of the most common ways AI has entered the daily life of anyone who writes, thinks, or makes things for a living. You generate something, then you immediately hand it to a system for judgment. The pattern is so familiar it barely registers as a choice. The question most coverage of this phenomenon asks is straightforward: was this idea really yours? It's a fair question, and it touches on something genuinely uncertain. Cognitive psychologists have studied this kind of thing for decades—researchers call it source monitoring, and the basic finding is that people are surprisingly bad at remembering where their thoughts came from. We routinely mistake things we read for things we thought, absorb an idea in conversation and later present it as original insight, lose track of the causal chain behind our own reasoning. The uncertainty you're feeling when you look at a piece of writing and wonder whether it was you or the tool? That uncertainty would exist even without AI. The question of "who generated this?" was never as clean as our folklore about creativity suggests. This is where the historical context matters more than most discussions acknowledge. The image of the lone genius—solitary, original, untouched by influence—is a relatively recent construction, a product of the Romantic era that got locked into our cultural mythology. Shakespeare borrowed plots. Beethoven worked within established forms. Scientists build on prior work in ways that are sometimes collaborative, sometimes competitive, always networked. The idea that a thought springs fully formed from a single mind is a story we tell after the fact, not a description of how thinking actually works. Creativity has always been more collective, more borrowed, more distributed across people and time than our mythology admits. But here's where the framing matters. All of that context—the source monitoring research, the history of collaborative creativity—tends to be offered as reassurance. "Don't worry," the argument goes, "your uncertainty is just a cognitive illusion; ideas were never purely yours anyway." That argument is not wrong. But it may be answering a different question than the one that's actually starting to matter. The more urgent question is not "was this mine?" but "who am I becoming?" Think back to the verification habit. What you are doing when you paste a draft into a chat and ask whether it's any good is not using AI to brainstorm, or to generate alternatives, or to catch your typos. You are using AI to tell you something about your own thinking that you could, in principle, tell yourself. The discomfort of uncertainty is real—nobody is pretending otherwise. But that discomfort is also information. It tells you something might need revision, or that you've made a choice you're not sure about, or that you've landed somewhere that feels exposed. The pause where you sit with your own work and decide whether it works is not a gap to be filled. It is the work. The capacity to sit with uncertainty about your own thinking is not a luxury. It is a cognitive skill, and like any skill, it responds to practice or disuse. This is where the agency question lives—not in the philosophical debate about whether AI-assisted work "counts," but in the practical question of what capacities you are exercising, and which ones you are letting atrophy. The concern is not that you are losing originality. The concern is that you are outsourcing the moment of judgment so consistently that you stop developing the capacity to judge. Notice I said "stop developing," not "lose." Agency is not binary. You are not either a person who has ideas or a person who doesn't. What you are doing, day by day, is practicing certain ways of engaging with your own thinking—or not practicing them. Each time you defer the question of quality to an external system, you are not making a catastrophic choice. You are making a small one, and making it repeatedly. The accumulation is what matters. This is also why the answer is not to put down the tools. That framing is too simple, and it doesn't survive contact with how knowledge work actually functions. The question is not whether to use AI—that ship has sailed, and more importantly, the question is not really about AI at all. The question is about the relationship you are cultivating with your own cognition. Are you still the kind of person who can sit with a draft and know, even roughly, whether it works? Not because you have outsourced that judgment, but because you have built the taste and the ear that allow you to make the call yourself? There are real tradeoffs here that deserve honest treatment. AI tools can do things that genuinely expand what you are capable of—generate variations you wouldn't have considered, surface assumptions you were blind to, compress time spent on routine revision. None of that is trivial. The problem is not the tools. The problem is the specific habit of delegation: not "help me explore" but "tell me if this is good." One expands your thinking. The other substitutes for it. The line between those two things is not always obvious in the moment, but it is worth being honest about which side of it you are on. A useful test might be this: can you articulate why something works, in terms that come from you? Not because you looked up a framework, but because you read it and something in you recognized it as right? If you can, the tool has been scaffolding. If you can't—if the best you can do is repeat what the system told you—then you have been practicing delegation, and you should ask yourself whether that is the direction you want to keep heading. The question worth sitting with is not "did AI help with this?" It is "am I still becoming the kind of person who can think through something, stand behind it, and revise it if needed?" That question is harder to answer than the attribution question. It doesn't resolve cleanly. But it is also a more useful one, because it is about direction rather than origin. It is about what you are building, not what you are tracing. And it is a question you can only answer by paying attention to your own habits—the small, daily choices about whether to sit with uncertainty or close it quickly, whether to trust your own taste or defer it, whether to stay in the uncomfortable but generative work of deciding for yourself what you think.