The Authenticity Trap

There’s a particular kind of discomfort that happens when someone compliments you on something you didn’t really do. Maybe it’s a cover letter that landed you an interview, one you fed your notes into ChatGPT and let it assemble into something polished and confident. Or a message to someone you were nervous to approach, where you asked the AI to “make it sound natural.” You got the response you wanted but somewhere in the back of your mind, a small uncomfortable question formed: did they just fall for a version of me I didn’t actually make?

Most people push that question aside. The outcome was good. Nobody got hurt. Everyone’s using AI nowadays. Why make it weird?

The question keeps coming back though. I think it keeps resurfacing because it’s philosophically serious. Not in an abstract, academic sense, but in the kind of way that eventually touches how you feel about yourself.

I suspect that Socrates would have found this moment fascinating because the problem it raises is exactly the one he spent his life prodding at.

The Delphic Oracle’s instruction, know thyself, was the animating principle of his entire project. Socrates likely understood something that tends to get lost in the self-help version of that phrase. Knowing yourself wasn’t a thing you could accomplish and file away. It wasn’t information to be retrieved. It was an ongoing process of uncomfortable, often inconclusive, and necessarily your own, inquiry.

That last part mattered to him enormously. The Apology makes this clear. When Socrates went around Athens questioning people, he wasn’t just being irritating for sport. He was genuinely alarmed by the number of people who held opinions about themselves, about their virtue, their wisdom, their values, that they had never actually examined. They’d absorbed these things from their families, their social class, the conventional wisdom of the city. They believed they knew themselves. They didn’t.

What he called doxa, received opinion, unexamined assumption, wasn’t knowledge. It was borrowed furniture that you’d stopped noticing was someone else’s. Here’s where AI enters the picture in a way that I think Socrates would have found genuinely troubling.

When you ask an AI to write your dating profile, something specific is happening. The AI produces a plausible, often quite good, version of how a person like you might present themselves. It draws on patterns across millions of human self-presentations and generates something fluent and appealing.

But fluent and appealing according to what? According to what tends to work. According to what is recognizable and legible to strangers. According to the aggregated preferences of people you’ve never met. In Socratic terms, the AI is essentially producing doxa. Not your examined self, but a kind of distillation of conventional self-presentation, and you’re signing your name to it.

Now, you might reasonably push back here. Humans have always had help presenting themselves. We ask friends to read our CVs. We hire career coaches. We take advice from people who know how these things work. What’s different about AI?

I think there are a few things, actually.

The scale is different. The AI isn’t drawing on the perspective of someone who knows you. Rather, it’s drawing on the preferences of the crowd, optimized for legibility and appeal. The result tends toward a kind of smoothed-out, frictionless version of personhood that has the shape of a self without necessarily having the substance of yours.

The ease is different too. The friction of writing your own profile including the false starts, the things you deleted because they felt too honest or too weird, the choices you made about what to include is actually doing something. It’s a form of self-examination. When you bypass it entirely, you bypass the inquiry.

There’s a Socratic dialogue called the Alcibiades (probably written by a student rather than Plato himself, but it captures something important) where Socrates tries to get the brilliant, ambitious Alcibiades to understand that he doesn’t actually know himself. Alcibiades thinks he does. He has a clear picture of his talents, his character, his values. Socrates keeps asking him to look more carefully, and the picture keeps getting murkier.

The point isn’t that Alcibiades is stupid. The point is that self-knowledge is harder than it looks, and that the version of yourself you carry around confidently is often quite partial. You’ve built it from habit, from feedback, from the stories you’ve been told about yourself. Most of it has never been tested by real inquiry.

AI, at its most convenient, gives you a way to present a self without having to know what that self actually is. It can write the profile before you’ve done the examining. The discomfort you feel afterward might be, in part, the Socratic alarm bell…the small recognition that something is being claimed on your behalf that you haven’t yet earned the right to claim.

None of this is an argument that using AI tools is wrong. That would be a much simpler essay, and also a fairly useless one. What it is, instead, is an invitation to take that discomfort seriously rather than dismissing it. The question did I actually write that? is, at root, a philosophical question about authorship, identity, and what it means to present yourself to the world. It has been a serious question for a very long time. The fact that it’s now being triggered by a chatbot doesn’t make it less serious.

If anything, the scale at which people are now outsourcing their self-presentation makes it more urgent. Not as a moral panic, but as a practical question: at some point, if you keep signing your name to things you didn’t really make, you may lose touch with what you’d make on your own. And that, Socrates would tell you, is a problem worth examining before it gets worse.