Creative Governance: A Compass to Powerful Authorship

(Beth Turnage Blog)

There’s an ongoing argument in writing communities about whether artificial intelligence can write good fiction. The usual objections are familiar. AI produces bland prose. It relies on predictable plots. It doesn’t understand human emotion. It smooths away unusual twists and turns in favor of conventional storytelling.

There’s some truth in those observations. Left to its devices, an LLM tends to produce statistically familiar results. It defaults to average writing, which was gleaned from human writers.(One it tempted to say, oops here.)

But there’s a fundamental problem with the argument.

Writing good fiction requires considerably more than producing good prose.

After years of professional fiction writing, ghostwriting, and developing outlines for commercial genre fiction, I’ve identified six interconnected elements that contribute to an effective story. Five concern the narrative’s construction. The sixth concerns the creative intelligence necessary to bring them together.

AI can assist with every one of these elements. But assistance isn’t authorship.

Authorship is the exercise of governing creative intelligence over a work. And that’s what gives our work value, no matter how we render it.

Let’s examine what that means in practice.

1. Wordsmithery: The Mechanics of Effective Prose

Most writers are familiar with advice about varying sentence lengths, avoiding passive voice, and eliminating unnecessary words.

But how many examine the measurable characteristics of their prose?

I’m talking about lexical density, pronoun percentages, weak verb percentages, average sentence length, and sentence-length variation.

These metrics can reveal problems that aren’t immediately apparent during ordinary editing.

Excessive pronoun usage can make prose monotonous or confusing. Too many weak verbs can drain energy from an otherwise exciting scene. Uniform sentence lengths can produce a mechanical rhythm that lulls readers when they should be engaged.

AI can identify these patterns and recommend changes. In fact, this is one of its more useful applications.

But metrics aren’t commandments.

A rapid sequence of short sentences might be exactly what a confrontation needs. Repetition might communicate a character’s obsessive thinking. A deliberately long sentence can immerse the reader in a character’s racing thoughts.

The human author must understand why a particular prose construction works before deciding whether to change it.

AI can measure the prose. The author determines its purpose.

2. Story Structure: Every Chapter Has a Job

A novel isn’t simply a sequence of interesting events.

It’s a carefully constructed progression of developments designed to produce a particular narrative and emotional experience.

Every chapter needs a job.

Does it establish a conflict? Complicate an existing problem? Reveal something essential about a character? Change the stakes? Prepare the reader for a later revelation?

And does it accomplish that job within the available word count?

A 30,000-word novella has different structural demands from a 100,000-word novel. You cannot allocate the same amount of narrative space to every subplot, relationship development, or worldbuilding detail.

AI can generate beat sheets, chapter outlines, and entire narrative structures. It can suggest where a reversal, revelation, or emotional turning point might occur.

But it doesn’t inherently know which events are essential to your story.

An AI might suggest removing a quiet conversation because it doesn’t advance the external plot. Yet that conversation may establish the emotional foundation necessary for a devastating decision six chapters later.

The author must understand the architecture of the entire work.

A chapter isn’t successful merely because something happens. It’s successful because the right thing happens at the right point in the story.

3. Worldbuilding: The Environment Must Affect the Characters

Worldbuilding isn’t limited to science fiction and fantasy.

Every story takes place within a world governed by physical, social, economic, cultural, and institutional realities.

A contemporary romance has a world. So does a police procedural. So does a historical mystery.

And good worldbuilding is more than describing the scenery.

Characters don’t simply move through their environments. Their environments act upon them.

Consider a character raised in a society where family obligations supersede individual desires. That cultural reality should influence how the character approaches relationships, loyalty, and personal ambition.

Or consider a society where professional certification is available only once in a person’s lifetime. That single institutional rule could transform an ordinary examination into a life-altering event.

AI can generate an elaborate fictional city, describe its political institutions, invent its history, and populate it with interesting people.

Wonderful.

But do those details matter?

Does the political system constrain the protagonist’s choices? Does the city’s geography complicate the investigation? Does the culture create conflicts that couldn’t exist in another setting?

If not, the worldbuilding is decoration.

The human author decides which details belong in the story and, more importantly, how those details influence what happens.

4. Character Psychology: Wounds Drive Decisions

This is where the distinction between generating character information and writing character-driven fiction becomes particularly important.

AI can construct an impressive character biography.

Give it a few instructions, and it will happily produce childhood traumas, personality traits, attachment styles, fears, desires, strengths, and weaknesses.

But a list of psychological characteristics doesn’t make a compelling character.

The character’s psychological history must influence the decisions that drive the plot.

Let me give you an example from my upcoming science fiction novel, Master’s License.

The protagonist is a fourteen-year-old girl named Kelleen Rynin. She’s a prodigy by almost any measure, raised aboard a spaceship by a strict father who has relentlessly schooled her in the knowledge and skills necessary to command a vessel.

Her intelligence and technical abilities are extraordinary.

But she’s still fourteen.

And she’s about to face a physically dangerous examination to earn her master’s license, the certification that will allow her to captain a ship.

Even Kelleen believes she’s too young for the test.

Unfortunately, the rules of her society allow only one attempt. There are no do-overs. If she refuses the examination or fails it, the opportunity is gone forever.

That’s the external conflict.

The internal conflict is considerably more complicated.

Kelleen’s father has spent her entire life preparing her for this moment. His demanding expectations have become the framework through which she measures her own worth.

She doesn’t simply want to pass the test because she dreams of commanding a ship.

She wants to please her father.

That motivation informs virtually every decision she makes.

She’s frightened, but she cannot allow herself to quit. She’s intelligent enough to recognize the dangers, yet emotionally compelled to proceed. She’s been taught to uphold certain standards, but when circumstances threaten her chance of success, she finds herself crossing moral boundaries that would ordinarily be unthinkable to her.

Her intelligence doesn’t protect her from those contradictions.

In some ways, it makes them worse. She’s clever enough to rationalize decisions that conflict with her own values.

The result is a character whose extraordinary abilities exist alongside the emotional vulnerabilities of a fourteen-year-old girl.

Her actions aren’t dictated solely by the requirements of the plot. They emerge from the psychological conditioning of her upbringing, her relationship with her father, and her desperate need for his approval.

And notice how the different storytelling elements intersect.

The world’s licensing rules create the external stakes. Her father’s demanding upbringing establishes her psychological motivations. Her technical education makes the examination possible. Her youth makes the experience emotionally and physically dangerous.

The story’s structure forces those elements into conflict.

AI could generate a backstory for a gifted young spaceship pilot. It could suggest that she has a demanding father, struggles with self-doubt, or wants parental approval.

But those are merely character details until they influence what she actually does.

The author must understand how those psychological forces interact and carry their consequences through the narrative.

Character-driven fiction isn’t about giving characters interesting histories. It’s about making those histories matter.

5. Audience Expectations: The Reader Promise

Of all the elements of commercial fiction writing, this may be the most frequently underestimated.

And in today’s highly segmented fiction marketplace, it may be the most important.

Readers don’t simply purchase books because they’re well written.

They purchase particular kinds of narrative experiences.

A contemporary reverse-harem romance reader has different expectations from a sports romance reader. A dark romance reader may expect emotional dynamics that would be entirely inappropriate in a sweet romance.

Even within a single genre, readers recognize increasingly specific combinations of tropes, conflicts, character dynamics, and resolutions.

These expectations form what I call the reader promise.

When a book is marketed within a particular subgenre, its presentation makes an implicit promise about the experience it will deliver.

A writer who misunderstands that promise can produce a beautifully written, structurally sound, psychologically sophisticated novel that nevertheless disappoints its intended audience.

I’ve seen evidence of this in my professional work.

For one contract fiction project, I developed an extensive guide to the conventions and reader expectations of its particular romance subgenre with AI assistance.

I then used that guide to construct simulated Alpha and Beta readers in AutoCrit and evaluated the manuscript through their perspectives.

Those simulated readers identified weaknesses in how the story fulfilled certain genre expectations.

I reviewed their observations, made editorial decisions, and revised the manuscript.

The resulting book performed exceptionally well in its intended market and received strong reader ratings.

Now, I can’t claim those revisions alone produced that success. There are too many other factors involved in commercial publishing.

But the experience reinforced something I’ve observed repeatedly in my ghostwriting career.

Understanding what a particular audience expects is a specialized writing skill.

In my professional outlining work, detailed subgenre guides and market-informed reader expectations are incorporated into the narrative architecture before the manuscript is written.

Different ghostwriters may subsequently render those outlines into prose, yet the resulting books generally receive favorable reader ratings.

That’s not proof of a universal formula for success. It is, however, practical evidence worth examining.

And it’s a reminder that an accomplished writer in one subgenre doesn’t automatically understand another.

AI can help research genre conventions, identify trope combinations, and simulate likely reader reactions.

But those simulations aren’t actual readers, and their judgments aren’t infallible.

The human author must evaluate the information, decide which expectations matter, and determine how to fulfill them without sacrificing originality.

Because fulfilling the reader promise doesn’t mean writing predictable fiction.

It means understanding which promises can be creatively challenged and which must be honored.

6. Narrative Integration: The Governing Creative Intelligence

Here is where everything comes together.

The first five elements don’t operate independently.

A decision that improves one element can damage another.

Consider a passage containing repetitive language.

At the wordsmithery level, repetition might appear to be a flaw.

At the character level, it might reveal obsessive thinking.

At the structural level, it might establish a pattern that becomes important later.

At the worldbuilding level, it might reflect a cultural belief or institutional practice.

At the audience level, it might contribute to the psychological intensity readers expect from the subgenre.

Should the repetition be removed?

There isn’t a universally correct answer.

The author must evaluate its function within the entire narrative.

Or consider Kelleen’s examination in Master’s License.

The test isn’t merely an exciting sequence of physical challenges.

It exists because of the rules of her fictional society. It matters because of her father’s expectations. It becomes dangerous because of her youth and the examination’s demands. It generates moral conflict because her determination to succeed sometimes overrides her better judgment.

And it provides the kind of high-stakes adventure and character development that readers of coming-of-age science fiction may expect.

Every element reinforces the others.

Remove the licensing rule, and the urgency diminishes.

Remove her father’s influence, and her motivation changes.

Remove her youth, and the psychological stakes become fundamentally different.

Remove the moral compromises, and the story loses an important dimension of her character development.

AI might propose any number of changes to those elements, individually.

But the author must understand what happens to the entire story when one of them changes.

Narrative integration is the exercise of creative judgment across all five disciplines simultaneously.

It’s the difference between assembling competent story components and constructing a coherent, emotionally satisfying novel.

So, Can AI Write Good Fiction?

I think that’s the wrong question.

AI can produce prose. It can generate outlines, invent worlds, construct character histories, analyze narrative structures, and research genre conventions.

It can also make mistakes in every one of those activities.

So can human writers.

The important question is whether the person directing the work understands the craft well enough to recognize those mistakes, make informed decisions, and maintain a coherent artistic vision.

An inexperienced writer can produce a weak novel without using a single AI-generated sentence.

An experienced author can use AI extensively while retaining control over every significant creative decision.

Neither the presence nor the absence of AI guarantees literary quality.

And typing every word yourself doesn’t automatically mean you’ve mastered the six elements of effective fiction.

There is a meaningful difference between asking AI to produce a novel and directing AI through a carefully conceived creative process. There is also a difference between directing a process and merely approving whatever the software produces.

That distinction matters.

The author must understand the story’s architecture, its characters’ psychological motivations, its world, its intended audience, and the language needed to bring everything to life.

The author must recognize when an AI suggestion is useful, when it’s irrelevant, and when it would actively damage the story.

Most importantly, the author must understand how all these elements work together.

AI can provide the materials. It can help analyze the construction. It can even perform some of the rendering.

But it doesn’t relieve the author of the responsibility to know what they’re building.

Authorship is the exercise of governing creative intelligence over a work. And that’s what gives our work value, no matter how we render it.