The Writer-Novelist In The World of AI

Are you a writer/novelist? An aspiring one? I’m curious about thoughts and threats (if any) for the fiction writer (and non-fiction, as well) in the looming world of generative AI. Is the creative writer/novelist about to become a relic?

My thoughts are that the risk is not limited to novelists, either, but to Graphic Artists as well…just faster, and perhaps already putting the designer out of the traditional business model. A junior artist could be hired to spend 10 hours rendering a specific fantasy landscape for a book cover. Today, a single prompt can generate ten unique, high-resolution versions of that landscape in 60 seconds. The market for generic visual execution is probably already gone.

But, I’m focused mainly on the Novelist, since I like to write and have trouble executing straight lines. Writing is certainly fun, but will AI wait around for the two year, two thousand hour average effort to turn out a completed novel, when it is able to generate the same 80,000 words in a minute, or two?

There is exactly zero art in creating a novel with AI, and I think art always finds a way to survive. Stolen words and no heart. I think all of this will hurt writers in the short term, but I also think long term readers will return to the art, which has meaning.

Perhaps I’m being overly romantic, or hopelessly optimistic.

Spicy hot take: If you are using AI to write, you are not a writer. You are a prompt engineer.

5 Likes

It is unlikely I can change how AI is muscling its way into the creative arts. But what I can do is resolve to never let machines write my novel series prose or the symphonic score I’m composing for it. The rest of it is outside of my control.

3 Likes

You’ve made a couple of excellent points here.

1, Surviving art: Exactly correct. Painting didn’t end, when photography came along.

2. Prompt engineering: I decided to test that very issue and quickly constructed a prompt for Gemini: A description of the floating US Senate building in the “Memory Keepers” epoch, highlighting the stark contrast between the lavish, golden exterior and the empty, automated hallways. The perspective should be from a low-altitude drone, emphasizing the building’s isolation and useless grandeur.

The Gemini response took 8 seconds: The drone, a cheap box of rattling sensors, clawed its way through the high-altitude haze, hugging the thermals below the Zenith Platform. Above, three hundred meters of cold-forged brass and synthetic pearl hung against the sky—the floating US Senate. It was a monument that defied gravity and good sense, a structure so determinedly golden it looked less like wealth and more like sickness.

The low, smog-choked sun hit the exterior panels, throwing back a glare that tasted of obsolete budget allocations. Everything about the design was meant to project permanence, to suggest that the Law was above the weather, above the dirt, above the people it was supposed to serve. But seen from this low vantage, it was just a sealed box, hermetically protected against the real world below.

Slipping through a defunct ventilation grid, the drone traded the sterile chill of the exterior for the hyper-conditioned silence within. The hallways stretched, automated and empty. Marble, polished monthly by mechanical routines, reflected nothing but distant servitor lights that tracked the drone’s progress—bright but utterly unconcerned. There was no dust. There was no life.

The central chamber was the cruelest joke. Mahogany and velvet, spotlighted perfectly. The tiered seating was immaculate, awaiting an audience that had long since dissolved into data packets or grave soil. Everything—the voting mechanisms, the translation booths, the automated gavel—waited.

It was the ultimate preservation project. The most expensive stage ever built for a play that would never open again. The Senate floated, maintained by silent protocols, ensuring that even in the final epoch of forgetfulness, this useless grandeur would remain aloft, a perfectly preserved memorial to the fact that when the world was dying, they spent their final credit on a gilded, empty shell.

Your thoughts are far closer to mine. I write for the love of writing, even if AI changes the world around me.

EDIT: I was/am highly resistant to AI composing a story for me. Yet, I recognize that the hurdle of ever getting new projects published may have just ratcheted up to a higher notch.

1 Like

It reads fine. It has no heart, or voice?

Bear in mind the LLM was likely (…elegegly?) trained on stolen content and is just pattern matching words. It all becomes pretty generic and ‘meh’, while ripping-off actual writers. It’s anti-art.

For full disclosure I use LLMs a lot for work (code, software engineering). There are good use cases, even in writing, but (to me anyway) limited to the technical aspects of writing. Not the words, as such. I’m a dyslexic writer and various writing tool are extremely helpful me keep prose technically correct (grammar, and my brain farting out the entirely wrong words when tired). But never, ever for composition. I never let Grammarly rewrite a sentence, for example.

I do use ChatGPT for some elements of research when hashing out ideas. But again, never prose. Never plot.

Edit: If someone switched off all AI tools tomorrow, I would still be able to write, and code. I worry there is a whole generation of folks being encouraged to lean hard on AI without ever learning the actual medium.

1 Like

Again, excellent points. Thank you. What if the prompt was longer, (a few pages, even) and included samples of your own writing – style, cadence, preferred metaphors, dialogue patterns- – so that your voice really did come through? What if you provided it with enough prompting material so that its output was clearly from your original thoughts and creativity? With enough detailed input, is it possible for the LLM to produce a finished novel that is truly original, copyrightable, publishable, with the soul that even I feel is missing from the meager prompting I provided? I still have creative input, but measured in minutes or hours, maybe days, but not in years. How about then?

1 Like

In another thread, I wrote:

Later, I was fascinated, purely as an amateur because I was now a lawyer, to follow the research into machines that can mimic human intelligence. I remember idly making notes (now in the eighties) about combining quantum mechanics with layered nodes to mimic neurons (an idea I had read about), but with the capacity to consider all neural pathways simultaneously (an idea that I had). This thought so frightened me that I have had a hard time enjoying the idea of AI ever since. Imagine a thinking machine that can consider all possibilities at once, while we humans are limited to linear neural pathways. If we think AI is too disruptive now, well, we haven’t seen anything yet. When the bastards finally figure out how to combine quantum computing with classical AI (it’s funny calling AI classical, but that’s what it is), we are really screwed as living, thinking beings. And I’m an optimist by nature.

My point is that we are only at the beginning of what AI can do in the realm of writing, limited, as it is, to large language models. Once we develop thinking machines based on qubit superposition, not only will the writing produced be indistinguishable from our best novelists, but it will surpass them because machine do not sleep, do not get tired, and never quit of their own accord.

So, each morning at 4 a.m. I pad down to my studio, bring up Scrivener and Aeon Timeline, and tap out my novel. It’s my way of shaking my fist at the bastards, one letter at a time.

LOL….I looooove your response. As I approach 79 next weekend, I feel like raising the same digital finger in their direction. Yet, I’m also a pragmatist, aware of what’s coming (or seems to be). I have no children to warn, but nieces and nephews who are already feeling the AI impact, and one particular niece, who at 25 is an amazing graphic artist. Her dad and I are already walking her down the “prompting-as-a-career” pathway, where she might continue to produce her own artwork, but more for her own edification. The message is that she may have to abandon the role of ‘executioner’ into the role of ‘director’.

1 Like

The question we’d have to answer is, what exactly are we presenting to the world as our work? We’ve written a nice, detailed, even amazing prompt–is that what we share, the prompt, and invite others to see what their LLM of choice can do with it? (I doubt it.)

Or do we publish the novel while disclosing that our role in writing it was merely creating the prompt? (I doubt it.)

Or do we pass off the “novel” as fully our own work, while happily collecting royalties as a fraud?

This is so sad. My grandchildren, who are toddlers, are inheriting a world that makes me want to cry.

Excellent observations and questions. But, what if the answer isn’t in any of the three you provided, but in a fourth? What if the prompt is so good, that I can sell it alone to a publisher, and let them take on the part that AI can already do so well…the final output? What if this is the exact future model for Intellectual Property ownership: the value lies not in the typing, but in the unique human vision that guides the machine. What if a Scrivener platform morphs from a place of full manuscript creation into a platform of really splendid and complete prompting creation/assembly?

1 Like

I feel the same. I was an actual radio DJ in my misspent youth and early adulthood, and I recall a similar feeling when fully automated (announcer-less) programming arrived.

1 Like

Great questions.

Let’s say that things (even Scrivener) arrive at a place where we provide the unique human vision* and let AI create the full work (say, a novel). We are still faced with how we (including the publisher) represent that novel to the reading public, are we not? Or are we okay with the public thinking we wrote the full novel ourselves? Will the public even care, so long as the story is entertaining?

*Machines that can mimic neural pathways could well be considered “human” for purposes of “unique human vision,” for they can, in theory, think like we do. Once we combine mimicked neural pathways with quantum particle superposition, a machine can consider every story possibility simultaneously; at that point, what hope will we have to provide anything unique that would not have already been imagined by the machine?

An excellent observation, but what if I sold away all future rights for a reduced price today on my splendid IP creation, and let the publisher make all of those decisions? If the alternative is spending the next two years of my life writing a finished manuscript, or banging on closed doors today trying to sell a completed manuscript, I may have made the better financial decision for my time. It becomes more of a pure business decision, but one that frees up a lot of my time, and the monumental risk I’m taking by choosing the path of writing a full manuscript. This naturally assumes that I don’t already have a publisher relationship and ready acceptance of what I write…which is likely the state of most would-be novelists.

1 Like

So, the idea is that a new profession of prompt writing rises to compete with the classic profession of novel writing. No doubt, that is happening; the economics of taking a massive shortcut to producing a novel are compelling, especially since machines will get even better at replacing even the best novelists.

But they shall not replace me, at least. I will continue to write and compose the old-fashioned way. Fortunately, I do not rely on writing income and have the financial freedom to walk the path of authenticity. Young, beginning novelists who aim for a career in writing, and who are not financially secure, may well find it irresistible to hand over much of their creative process to the machines.

Spot on, @SCN. I’m fortunate enough not to have completely squandered earlier earnings on too many fancy trappings. I had hoped and planned that this day would arrive. So, I don’t write for income, either, and it’s a good thing. I have three completed novels: One self-published after meeting the hard publisher’s wall head-on. Two that are complete, but not even sent to DC for copyright. They are sheer personal endeavors. And, there are six more in various stages of completion, from nearly finished to ‘just an idea’. I call these “shelf titles”.

With 4-way bypass surgery still fresh in my rear-view mirror, and the available play clock time beginning to dwindle, I realize that even writing these six is likely impossible. Yet, I don’t necessarily want them to die with me, especially the complicated one….the one where I still need to find a way for sodium-ion ray gun in the hands of the evil Xlara Tribe Commander, to magically disappear and then reappear in the hands of my protagonist, so that he can reawaken the love of his life..who has been put under a spell by the Witch of Watoon, who lives on the small red moon, Blorda7, orbiting Xlara. :scream: :rofl:

That takes time and energy that I’d now rather spend on my significant other. You get the drift. It was with these thoughts in mind that I began to wonder how to keep these “amazing feats of narrative prose” alive, even after they’ve poured me into the waiting urn. And, that all brought me to the realization that my keyboard will likely out-write me in the near term, anyway. :rofl: ….which then led me to my Scrivener account, and to you fine folks this morning.

For my small group of writing friends who simply love the art of writing, I say, “Keep doing it. Write to your heart’s content. I wish you all the luck I never experienced. I hope to see your name in lights, or on the big screen.” But, to my friends who definitely write for income, I say, “The world is changing before our eyes. Let’s look for Plan B.”

1 Like

You’ve made my day! :laughing:

:rofl: Excellent. My work here is done. Thanks for all of your input. Very helpful, indeed.

1 Like

That’s correct….most especially about harnessing the power of AI where it can be useful, but not destroy the human creative element. Perhaps one day, my Scrivener Binder may have this outline (Composed with the help of Gemini)

Scrivener Binder: Project PROMPT-CEPTION (The Blueprint of the Blueprint)

This is the assembly file for the 100,000-word System Instruction that makes the LLM believe it is actually me. All documents here are intended for machine consumption only.

1. :open_file_folder: THE INSTRUCTION MANUAL (System Prompt)

  • :page_facing_up: 01.0 - Prime Directive: The core rule: “I am SkyPilot. You are my typewriter.”

  • :page_facing_up: 01.1 - The Legal Disclaimer: Why I own the output, not the crane operator.

  • :page_facing_up: 01.2 - The Emergency Stop Clause: What to do if the output starts sounding like a self-help book or, worse, uses “adjective-heavy flowery prose.”

  • :page_facing_up: 01.3 - Output Delivery Format: Instructions for delivering the text in Scrivener-ready .RTF format with zero typos (A girl can dream).

2. :open_file_folder: THE STYLE FINGERPRINT (Cloning SkyPilot)

  • :page_facing_up: 02.0 - Cadence & Rhythm (The Metronome): Detailed analysis of sentence length variation and where to place the surprising, short statement for maximum impact.

  • :page_facing_up: 02.1 - The Metaphor Library: Explicit instruction to use analogies involving quill/inkwells, obsolete machinery, or piano tuning at least once per 5,000 words.

  • :page_facing_up: 02.2 - Dialogue (The Northwest Arizona Accent): Rules for direct, thoughtful, and slightly wry character conversations. Avoid melodrama at all costs.

  • :page_facing_up: 02.3 - The Forbidden Word List: (e.g., “Synergy,” “Ideate,” “Pivot,” and any sentence beginning with “As we move forward…” or “Meanwhile”)

3. :open_file_folder: WORLD LORE & PHILOSOPHY (The Codex Humana)

  • :page_facing_up: 03.0 - The Volktra Principle (Final Definition): The concept of controlled, integrated darkness.

  • :page_facing_up: 03.1 - The Four Epochs: Brief, hyper-specific notes on what key technology/social dynamic defines each era (Memory Takers, Keepers, Finders, Reclaimers).

  • :page_facing_up: 03.2 - Setting Guide: US Local (Bella Norte Detail): A five-page document describing the exact look and feel of the environment, ensuring geographic accuracy down to the light reflecting off the nearest lake.

4. :open_file_folder: CHARACTER ARCS (The DNA)

  • :page_facing_up: 04.0 - Protagonist (The Uncanny Valley Human): Specific details on motivation, internal monologue, and musical taste (prefers minor keys).

  • :page_facing_up: 04.1 - The Antagonist (The Bureaucratic Vacuum): Must be a force of cold, emotionless efficiency, not a mustache-twirling villain.

  • :page_facing_up: 04.2 - Supporting Cast: Detailed background for the three characters who will appear for only one scene each, ensuring they have depth.

5. :open_file_folder: THE FORBIDDEN DRAFTS (The Trash Can)

  • :page_facing_up: 05.0 - Draft 1 (The AI went rogue): Initial attempt where the LLM wrote a story about friendly space dogs. (Locked and hidden.)

  • :page_facing_up: 05.1 - Draft 2 (The Philosophical Abyss): Where the text became too dense and stopped being a thriller.

  • :page_facing_up: 05.2 - The “Quill and Inkwell” Origin Prompt: The conversation that started it all.

6. :open_file_folder: FINAL ASSEMBLY (The Execute Button)

  • :page_facing_up: [MAIN COMPILATION FILE]: This is the document where Sections 1 through 4 are concatenated, reviewed, and prepared for copy-pasting into the LLM’s chat window. It is the finished, monolithic prompt.

  • :page_facing_up: The Final Scene’s Question Mark: The single line of code that ensures the ending is a perfect, existential cliffhanger.

1 Like