Lost in Translation – Jean-Jacques Rousseau Predicted the AI Profile Predicament in 1762.

On amour de soi, amour propre, and the strange new ways we perform ourselves for strangers.

There’s a conversation happening right now about AI and authenticity, and it mostly sounds like this: Is the content “real”? Is the person “genuine”? Are we losing something important?

These are the valid questions. The philosophical framework people reach for, however, is usually pretty thin…some vague gesture toward being yourself, maybe a mention of Sartre if they’re feeling ambitious. Nobody is talking about Rousseau, which is surprising, because Rousseau essentially invented authenticity as a moral concept.

The two kinds of self-love

In Emile and the Discourse on Inequality, Rousseau argues that humans have two fundamental modes of relating to themselves. The first is amour de soi, love of self, in the literal sense. This is the natural, pre-social impulse toward your own wellbeing and preservation. It’s not selfish in the pejorative sense. It’s attending to your needs, caring about yourself, existing as a coherent person with desires and preferences that are actually your own.

The second is amour propre, which is much harder to translate cleanly, and that’s kind of the point. It’s usually rendered as vanity or pride, but those words are too small. What Rousseau means is closer to: the desire to be seen in a certain way by others, which gradually replaces your actual self-perception as the source of your self-worth. It’s the version of self-love that can only be satisfied through comparison, through approval, through being perceived correctly.

For Rousseau, amour propre is a corrupting force, not because caring what others think is inherently wrong, but because it creates a feedback loop where being a person is replaced by being a performance. You lose track of what you actually want because you’ve been optimizing for what reads well.

Now look at your phone

Whatever app currently has the most culturally anxious think-pieces written about it. You upload photos, you answer prompts, you write a bio. The whole thing is, structurally, a performance for strangers. Rousseau would likely recognize this immediately as a context where amour propre runs the show. You’re not expressing yourself; you’re constructing yourself for external evaluation. The question of what you actually like, what you actually want, what you actually sound like, all of that gets filtered through: but will this be received?

This is already a tension that existed before AI. People have always workshopped their bios with friends, hired profile consultants, read Reddit threads about which photos perform best. The performance-for-approval problem isn’t new, but AI does something qualitatively different.

When you hand any online profile to ChatGPT or Claude prompting it to make improvements, what you get back is a profile optimized for legibility, warmth, and appeal. It’ll probably be better as a piece of writing than what you’d produce yourself. That’s precisely the problem Rousseau would identify.

Amour propre, in Rousseau’s account, isn’t just about vanity, it’s about the substitution of the external gaze for the internal one. When you let an algorithm (or a model trained on millions of humans) generate your self-presentation, you’re not just polishing your amour propre. You’re outsourcing it entirely. The self being performed for others’ approval is no longer even recognizably yours, it’s a statistical aggregate of what tends to read as charming.

There’s a difference between I wrote something that performs well and something that performs well was written in my name. The first involves some negotiation between your genuine self and the social context. The second skips the negotiation. Rousseau would say: the negotiation is the authenticity. The awkward, effortful process of figuring out how to present yourself honestly, that’s where you find out what you actually value about yourself, what you’re willing to claim, what you’re proud of. Short-circuit that process and you don’t just get a better bio. You get a bio that has no relationship to amour de soi at all.

A wrinkle worth reflecting upon

Rousseau himself was not exactly a paragon of authentic self-presentation. His Confessions, the book he wrote specifically to present himself truthfully to the world, is famously selective, self-serving, and in some places demonstrably false. He wanted to be seen as authentic more than he wanted to be authentic, which is a very human problem and arguably an amour propre trap he couldn’t escape himself.

So, there’s a version of this argument where write your own profile is just another performance, another way of signaling I’m the kind of person who does things authentically, rather than an actual expression of amour de soi. I think the distinction still holds at a structural level, even if the execution is always messy. The question isn’t whether your self-presentation is pure and unmediated (it never is). The question is whether there’s a you in there making choices, taking responsibility, deciding what to claim. AI-written profiles remove that agent from the loop in a way that feels categorically different from asking a friend for feedback.

What Rousseau would recommend

Rousseau was generally pretty down on modern society. Within the frame of the problem: he’d push toward anything that reconnects the self-presentation to actual self-knowledge. Write the messy first draft. Leave in the specific detail that’s hard to explain but true. Don’t optimize for relatability, optimize for accurate, and trust that the right person will find accurate more appealing than legible.

The irony is that genuine and slightly awkward tends to work better on profiles anyway. People can feel the AI smoothness. It’s like the uncanny valley, but for personality. Rousseau’s real insight isn’t that performing for others is evil. It’s that when amour propre fully displaces amour de soi, you lose the ability to evaluate yourself on your own terms, which makes you chronically dependent on external validation in ways that aren’t actually satisfying. To use online dating profiles as an example, the person who matches with your AI-generated profile is, in some sense, matching with a fiction. The subsequent date is going to be a negotiation between the fiction and reality, and that’s a weird place to start something. Online profiles are just the most legible example because the stakes are clear and the dynamic is obvious. The same framework applies to AI-assisted cover letters, LinkedIn summaries, performance reviews, essays, really any context where you’re constructing a self for external evaluation and an AI can do it more smoothly than you can.

The alienation problem

The dating profile and cover letter are just the most visible examples of something that goes deeper. What Rousseau’s framework actually predicts, though he couldn’t have named it this way, is a specific kind of self-alienation that becomes possible when you have a tool capable of generating your self-expression on demand. The classic concept of alienation is about being separated from the product of your labour: you made something, but it doesn’t belong to you. This is structurally similar but more intimate. You become separated from the product of your self-expression. The profile, the essay, the cover letter, these things exist and go out into the world and produce effects, but the causal chain between your inner life and those effects has been severed. Something was made in your name that didn’t come from you.

That might sound like delegation, and we’ve always delegated, ghostwriters, editors, friends who read a draft and tell you it doesn’t sound like you. It’s just that there’s a compounding effect worth taking seriously. When you consistently outsource the process of self-expression, you stop practicing the thing that self-expression is supposed to develop. That process isn’t just a means to an end; it’s how you find out what you actually think. Writing the clumsy, honest first version of something is how the honest version becomes available to you at all. Skip that process enough times and you don’t just lose the artifact, you lose access to what it was pointing toward. You stop knowing what you’d say if nobody were helping.

This is where it connects to meaning and purpose, and it’s worth being precise about the mechanism. A life that feels meaningful tends to be one where you can trace a line between your values and your actions, where what you do and say in the world feels like it actually comes from you, rather than from a version of you optimized for legibility. There’s a thread running from Rousseau through the existentialists that insists you can’t borrow your way to a purposeful life. The sense of meaning that comes from having made something, said something, taken a position, it requires authorship. Not in the legal sense. In the deeper sense that the thing genuinely came from you: your judgment, your willingness to be seen, your decision about what to claim.

When AI mediates more and more of that expression, the line becomes harder to trace. Not because your values have changed, but because the translation layer between your inner life and what you put into the world has been quietly handed to something else. Over time, you might find yourself in the strange position of living a life that looks like yours – your job, your relationships, your words on the page – but that feels somehow adjacent to you rather than fully inhabited. That’s not a dramatic collapse. It’s more like a gradual dimming. Rousseau’s fear was that amour propre would make you dependent on others’ eyes to feel real. AI introduces something a little different: a version of yourself that is smoother and more presentable precisely because someone else assembled it. The cost is that it never quite belongs to you, and on some level you know it.

In each case, Rousseau’s question is the same: whose amour propre are you feeding? What is the relationship between the document and the person? And, most uncomfortably, have you spent so long optimizing your self-presentation for others that you’ve genuinely lost track of what you’d say if nobody was evaluating you?

The AI-facilitated alienation is almost a symptom. The disease is the same one Rousseau diagnosed in 1755: we’ve built a society that makes amour propre almost impossible to resist, and the tools just keep getting better at helping us perform.