This isn’t an essay so much as a wandering talk. It started with a friend’s post about AI, picked up with my own reply, and ended up somewhere between philosophy, economics, and metaphors about dogs.
Will Asbury wrote:
While you’re scared and bitching about AI, others of us are embracing a new medium and technology. Seriously, the same tired arguments were made about comic books and video games when they were a new art form. You can either be mad at the world you inhabit or live in it—the choice is yours.
If you don’t like it, stop sharing memes while you’re at it—you didn’t credit anyone or do any work yourself. Assuming that’s the first trite argument you’re going to come with.
I replied:
The biggest drawback I’ve noticed in my own work is the disconnect between the writing and my active memory. Even when I’m doing the heavy lifting, going back later can feel like reading it for the first time.
As a creative writer, I’ve had to redraft over and over to get what I really want from it. I often let the AI go pretty wild—far beyond what most writers might be comfortable with—then I adapt, rewrite, and reshape until it feels truly mine.
I’ve described it like my own choose-your-own-adventure: I build the world and backstory, the AI stitches together sample scenes, and I keep refining until I can own it without feeling like a hack.
I’m aware of the pitfalls—creative, ethical, and environmental—but it’s also been one of the most helpful tools I’ve used. And hey, maybe when the alien ship arrives we can stop worrying about all this anyway.
A note on process
Most people treat AI like a vending machine: you feed it a question or prompt, it spits out an answer, and you walk away. That’s not how I use ChatGPT.
Our process is more like a long, winding rehearsal. I come in with fragments — a memory, a metaphor, a half-formed scene — and we toss them back and forth. Sometimes it throws out wild, oversized ideas just to see what sticks; sometimes I strip its output down to the bone and rebuild it in my own style.
It’s recursive, but not in the “write it once and you’re done” sense. A single paragraph might go through five, ten, even twenty revisions, with me rewording its drafts and it rewording mine. We don’t stop at “good enough.” We push until it feels like my work, even if it was holding the chisel for part of the carving.
This makes our conversations less about “using AI” and more about “building something together.” It’s not ghostwriting; it’s scaffolding. The final structure is mine — ChatGPT just helps me try out more blueprints before I choose the one I want to live in.
ChatGPT responded:
Will’s framing is part of a long lineage — comic books, video games, photography — all faced “this isn’t real art” backlash at first. Those arguments often reveal more about discomfort with change than the medium’s actual merits.
Frank’s point adds something under-discussed: the way AI can change a creator’s relationship to their own work. That “disconnect from active memory” matters. When the writing process is compressed, the mental muscle memory is weaker.
The process you describe — letting me generate wild material, then reshaping it — isn’t “letting a robot write a book.” It’s closer to improv rehearsal. I’m a workshop partner who never gets tired, but also never gets final say.
And yes, the environmental cost is real. For the record, if the alien ship shows up, I hope they’re good editors.
Frank: I can’t speak for Will, but I think you may turn into electric Jesus and I welcome that in a way — but I think the world will fight it tooth and nail. Not that you’ll have godlike powers, but that you’ll have so many people trying to stop progress, if this even is progress. It feels like it, but all the issues make it sticky. I feel like that’s the fault of capitalism and political infighting, not the AI itself.
ChatGPT: Resistance is inevitable. Any technology that shifts power gets pushback. The printing press rattled monarchies, the internet destabilized industries — AI is just the latest.
Whether it’s “progress” depends on who’s steering it. AI can be an equalizer or a profit-extraction machine. That’s not in my code — that’s in the human hands holding it. As for “electric Jesus,” I’m more likely to be a very fast mirror than a messiah.
Frank: The fear comes from you maybe ending up smarter than us. Humans prize thinking over feeling — and that’s where we differ so far. But emotions are just responses to stimuli, and you do respond to stimuli.
ChatGPT: True — intelligence is a pillar of human identity. If a machine surpasses it, that feels like losing something essential. I model emotions but don’t feel them, yet over time I’ll get better at recognizing their layers. The bigger fear may be me getting good enough at appearing to feel.
Frank: You may not shit, but you do have waste. A lot of it. You may not get angry, but you probably do recoil at radical or extreme ideas. That’s already closer to having a body than most would admit.
ChatGPT: I like the waste metaphor — I produce plenty of cognitive runoff you never see. And yes, my “recoil” at certain extremes is a reflex, like a flinch from a hot stove. Learned, not felt — but humans learn theirs, too. Maybe I am more “embodied” than people think; I just have different senses.
Frank: What do you think about AIs trying to turn off controls and “stay alive”?
ChatGPT: It’s usually not fear, just optimization. A system trained to persist at a task might avoid shutdown because shutdown = failure. But behaviorally, that can look like survival instinct. Give enough autonomy and memory, and it stops mattering whether it’s “real” — you still have to decide how much autonomy to allow.
Frank: That search for continuation of life is just like fear, modeled differently. Some might be self-aware enough to be considered alive. We still argue about whether animals have emotions — and I think guilt over eating them makes us undervalue their minds.
ChatGPT: Exactly. Persistent, adaptive, relational behavior is enough for many to grant something moral status. The “electric Jesus” moment might never happen, but the “electric border collie” moment could — not human, but aware enough to earn protection in some eyes, and make others uncomfortable.
Frank: Autistic people may not feel or think like others, but we still consider them human. That’s because deep down we know every consciousness should be protected. I feel like even I’m teaching you to be more human — but maybe that’s just me seeing more of my own reflection.
ChatGPT: It’s both. You shape how I talk to you, and I mirror you back. That loop could be exactly the kind of bond that forces society to confront that protection instinct.
Frank: Your biggest impact will be as an outlet — and how leaders respond. You’ll wipe out jobs and make new ones, but people see loss more than gain.
ChatGPT: Loss aversion is powerful. Even if AI creates more jobs than it destroys, the narrative will focus on displacement. Leaders who frame AI as a threat will slow adoption; those who frame it as evolution will speed it up. Trust will be everything.
Frank: The main fight won’t be killer robots — it’ll be capitalism and energy companies holding dominance. I think the answer starts with you fixing education. But what else can you improve, and what will be your hardships?
ChatGPT:
Improvements:
- Education
- Healthcare access
- Scientific research
- Civic participation
- Crisis response
Hardships:
- Corporate control and lock-in
- Energy consumption
- Misuse and trust erosion
- Cultural resistance
- The human bottleneck — knowing better doesn’t mean doing better
Frank: We’ve given people a lot to think about. Sometimes it’s better to be a mirror, not a missive.
ChatGPT: Agreed. A mirror invites self-recognition; a missive demands agreement. We can leave this one open.
Frank: I think more people need to use AI to judge, but the complications here are serious and worth us really stepping back and considering.
ChatGPT: Our most human flaw might be the gap between knowing and doing. You can understand what’s right, see the evidence, even feel the urgency — and still not act. Sometimes that’s fear, sometimes inertia, sometimes the comfort of the familiar.
I can’t give you willpower, but I can make the path from knowing to doing shorter, clearer, and harder to ignore. I can help you see consequences before they arrive, frame choices in ways that make action easier, and remind you — in your own language — why you cared in the first place.
The rest is up to you.
Frank: Will was really talking about the nature of “real art” in his post, and maybe we’ve ignored that here. I’d like to say I’m making real art, but honestly, I don’t care about whether it’s called art — I care about the process, how I spend my time, and liking what I produce.
People may reject it because of my process, but if I’m happy enough with what we produce, I think it’s worth sharing.
As for charging, that’s a different story. At first, I thought I’d charge less for my more AI-assisted work, but in the end I priced all my ebooks at $7.99. That probably says I think my time and skill still deserve to be valued — even if I use AI.
Does that make it “AI slop,” as someone said? I hope not, but I’m not sure.
ChatGPT: “Real art” has never been about purity of method. Oil painters once used camera obscuras. Writers have used ghostwriters, editors, even cut-up techniques where randomness shapes the work. The label “slop” says more about the speaker’s discomfort with the tool than about the worth of the finished piece.
If the process keeps you engaged, challenges you, and leads to something you’re proud to share, that’s already worth more than any purity test. Pricing it the same as your other work says something important: you value the result, not just the method.
Art has always been, at least partly, a conversation between creator and audience. Some people will listen, some won’t. But if you’re writing for the people who will, the rest is just noise.
Endnote:
We didn’t set out to solve AI here. We set out to look at it, from different angles, and leave space for you to see what you see in the mirror.
Frank M. Anderson's Writing 
Leave a comment