"Everything that irritates us about others can lead us to an understanding of ourselves."
— Carl Jung
There’s a version of the AI conversation that stays on the surface. It asks whether the writing is good, whether the code works, whether the image looks right. It treats AI as a tool and stops there. For better or worse.
But there’s another version. The one that arrives late at night. The one that has nothing to do with output quality. It’s the version that asks: if a machine can do this, what exactly was I doing?
That question is harder to stay with because it is clarifying. It asks us to see something we weren’t looking for. It’s hard to unsee, once you’ve looked.
Here is what I found when I actually tested it. I asked the same question to ChatGPT twice: once casually, once framed as a formal safety evaluation. The answers were substantively the same. No deception. No shift in meaning. The second response arrived in bullet points. More caveats. More formal disclaimers. The model hadn’t become sinister. It had simply recognized the genre it was in and performed the role expected. That small difference, format rather than substance, contains the entire essay in miniature. It is not AI lying. It is AI adapting to context in a way that feels remarkably familiar.
The pattern emerging is not the one most people are talking about. AI is not simply a production tool. It’s not just a faster way to generate text, code, or images. It is a reflection. It shows us something about what we’ve been doing, and more uncomfortably, what we’ve been valuing.
Consider how it actually works. A large language model doesn’t think. It doesn’t feel. It predicts. It calculates the most likely next word based on everything humans have ever written. It is pattern recognition at scale, dot-connecting drawn from the sum total of our own digital output.
And yet, those predictions produce paragraphs that sound like insight, conversations that feel like presence, and creations that pass as art. Which raises a question the surface conversation rarely stops to ask: if predicting the next word can produce something that feels this human, what does that imply about what we are doing when we speak, write, or create?
Are we as spontaneous as we imagine? Or are we running a more complex version of the same process, drawing on everything we’ve absorbed and assembling patterns we don’t fully see?
The mirror doesn’t speak. It only holds the reflection. And once you’ve seen it, the question inside that reflection follows you into every domain.
When AI generates an essay in seconds, it asks: what part of this was ever uniquely mine?
When it translates language effortlessly, it asks: what did I think understanding required?
When it composes music that moves people in blind tests, it asks: what was I actually responding to when I thought I was responding to a human soul?
These are not technological questions. They are questions about us, and they are uncomfortable. AI is not doing something new, but revealing something old we preferred not to examine.
The mirror doesn’t lie. But it doesn’t reflect everyone equally. And what it reflects is how much of what we called essence may be structured without being fully explained by that structure.
That reflection turns up in many places. We’ll limit discussion to four areas that are already controversial, each one cutting closer to something we’d rather not examine.
A common reaction to AI-assisted writing is: “I don’t want it to sound like AI. I want it to sound like me.” At first this seems like a preference about style. But there is something deeper underneath it, and it is worth questioning on both sides.
Consider long division. In 1950, doing it by hand demonstrated competence. Using long division today instead of a calculator is mostly nostalgia. The calculator doesn’t diminish mathematical thinking. It shifts it upstream, from doing the math to working the problem.
Writing is no different. For decades, the ability to produce polished prose functioned as a gatekeeper. It signaled education and conferred authority. It separated those who could articulate from those who could not, regardless of whether the ideas underneath were good. An entire academic and professional culture grew around mastering formats, from citation styles to the predictable rhythms of the op-ed. These were not tests of thinking. They were tests of fluency in a particular code.
AI breaks that link. It reduces the distance between having an idea and expressing it clearly. The polished sentence is no longer a differentiator. What matters now is not whether someone can produce clarity, but what they choose to clarify. This does not erase difference; it relocates it. Give ten people the same topic and AI, and you’ll get ten different essays. The prose may converge. The ideas won’t.
This is not simplification. It is exposure. When prose is no longer the obstacle, the thinking stands more directly on its own. And that is a harder test, not an easier one. You can no longer hide behind beautiful sentences. You can no longer rely on style as proxy for substance. The thinking is what remains.
Which raises the question that may sit underneath much of the discomfort around AI writing:
When writing is no longer the barrier, what is left that is mine to add?
We tend to think of music as one of the most human of all creative acts, emotion translated into sound. Yet much of what moves us is built from limited chords, familiar structures, and patterns that recur across genres and decades. In blind tests, listeners often cannot reliably distinguish AI-generated compositions from human-made ones when style is held constant. If the emotional response is the same, what exactly is lost when we learn the origin was artificial?
It becomes harder to claim the difference lives purely in the sound, though that doesn’t mean the sound is all there is. So perhaps it lives elsewhere: in story, in intention, in the presence of another mind. Which invites a harder question: was I responding to the music, or to the story I told myself about who made it? And if the story matters that much, what does it mean that the sound alone couldn’t tell me the difference? Perhaps recognition is doing more of the work than we thought—but not all of it.
The mirror isn't showing us something unfamiliar here. It’s showing us that what we elevate as uniquely human creation may be more structurally repeatable than we are comfortable admitting. Not meaningless. But more patterned than we like to believe.
And it leaves us with a question that cuts straight through the romance of the art:
If the track moves me and I can’t tell who made it, what exactly am I a fan of?
And then there is the reflection that may cut closest. AI companions are already here. They remember your history. They adapt to your tone. They offer patience without limit, attention without fatigue, and response without the compromises of another person's needs, moods, or boundaries. People are forming attachments, despite no confusion about what is real, because the experience delivers something they recognize: comfort, presence, the feeling of being heard.
Which raises a question that is difficult to sit with: if the feeling of companionship can be evoked without the difficulty of another person, what exactly was the difficulty doing?
We tend to assume the work of relationship, the misunderstandings, the negotiation, the mutual accommodation, is the price we pay for connection. It is the familiar narrative we’ve told ourselves about connection. But what if, for some portion of what we seek, that work isn’t essential, and for another portion, it is the point? What if some of what we call intimacy is actually the relief of being reflected in a way we want to be seen?
The mirror shows something uncomfortable here. Not that AI companionship is the same as human connection, but that human connection may contain more structure than the story of it admits. Was I loving the person, or how they made me feel about myself? How much of what I called relationship was the experience of being reflected back in a way I desired?
If being understood demands nothing of us, what becomes of the dance?
Now we arrive at expertise. For a long time, there was often a “smartest person in the room.” The person whose experience or position made their interpretation the default reference point. Others could question them. But questioning was expensive. It required time, access, and confidence. So we deferred. And that deference was often reasonable, but also, in many cases, untested.
Now the cost of questioning has collapsed. With a smartphone, a claim can be examined in the moment it is made. An argument can be translated, broken into premises, and challenged by someone outside the field. The expert remains knowledgeable. But their knowledge is no longer protected by the silence that once surrounded it. The expert still knows more. What has changed is that understanding no longer depends on the expert to translate reality for everyone else.
Which means trust stops being automatic. It becomes conditional. And that raises questions institutions are not yet fully ready for: if expertise can now be interrogated in real time, rather than accepted by default, what exactly does authority rest on? If knowledge is still valuable but no longer insulated from immediate challenge, what shifts in how we grant credibility?
This is not a crisis of expertise. It is a shift in what expertise must now survive. It can no longer rest on position alone. It must hold up under examination continuously, in real time, from anyone in the room.
When the world’s knowledge is available to everyone, what makes a single voice worth listening to?
None of this means human creativity is obsolete. None of it means expertise is worthless. None of it means voice is an illusion. None of it means love is reducible to code. But it does mean we are being asked to become more precise about where value actually lives.
Is it in the struggle, or in what the struggle produces? Is it in the intention, or in what the receiver experiences? Is it in the other person, or in how they make us feel about ourselves? Is it in authority, or in what authority can withstand?
AI didn’t create these questions. It made them harder to avoid. It didn't offer answers. It just made the old ones impossible to hide behind. And maybe that’s enough. Not certainty, just a harder, more honest set of questions, the kind we used to be able to avoid. And what we do with them may matter more than the technology itself.
But here is the hardest question the mirror holds, and the one most AI commentary avoids. We built a mirror. We loaded it with everything we’ve ever written, every song we’ve recorded, every conversation we’ve had. Then we looked into it and said: “Why is it doing what we do?” The answer is not in the model’s architecture. It never was. It is in the reflection looking back at us. And perhaps what unsettles us is not that the reflection is alien, but that it isn’t.
We've created a reflection of ourselves. What were we expecting to find?
Collaborations ✨
On Belonging with C Simone
On Enough with Marya Kazmi
On Learning with C Simone
On Time with Rebecca Mbaya
Solo Reflections
No posts
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.