When Creative Assistance Has a Price

Ghostwriters and editors have long helped authors produce books, sometimes contributing substantially to the finished work. AI is making some of that assistance available to writers who could never afford it.
Francine Pascal created Sweet Valley High, the enormously successful young adult fiction series that began in 1983. She developed the characters, established their fictional world, and constructed the stories that followed Elizabeth and Jessica Wakefield through their increasingly complicated lives. The series eventually expanded into hundreds of books, including numerous spin-offs, and reached millions of readers. Pascal was responsible for the stories and their direction, while a team of ghostwriters produced the manuscripts from her outlines and instructions.
In a 2012 interview with The Guardian, Pascal described creating the characters and stories, then working with ghostwriters who transformed her outlines into novels. The division of labor was considerable. Pascal determined what happened within the fictional world, while the writers constructed scenes, developed dialogue, and made the countless smaller decisions involved in telling those stories. An outline might establish that two characters have an argument, but someone still has to determine how that argument unfolds, what the characters say, and how the scene changes their relationship. Those decisions require creative judgment, even when the person making them is working within a story someone else conceived.
Pascal retained creative control, and the books were associated with her name. The ghostwriters contributed substantial creative labor, yet their role didn't prevent Pascal from being recognized as the creator of Sweet Valley High. Publishing had a working arrangement for distributing those responsibilities, compensating the writers, and presenting the finished books to readers. The division between developing a story and composing its prose was understood well enough that the series could continue through the work of multiple writers without changing the identity under which it was published.
Now consider someone attempting something similar without the resources to hire ghostwriters. They create the characters, develop the plots, prepare detailed outlines, and use a language model to generate preliminary scenes and chapters. They review the output, reject material that doesn't fit, rewrite passages, and determine the direction of the finished work. Depending on how much they contribute, their role could be similar to Pascal's, considerably more involved, or much less so. The comparison becomes interesting because ghostwriting is an established publishing service, while using AI to perform comparable tasks can attract suspicion. We can examine the work each author performed and reach different conclusions about their creative contributions, but knowing that AI was involved doesn't establish how much of the story the named author actually created.
The Work Behind the Name
Most readers understand that books undergo editing. A manuscript may be corrected for spelling and punctuation, checked for consistency, or revised in response to suggestions from an editor. Few people would question an author's contribution because someone corrected their grammar, and ordinary editorial assistance is generally accepted as part of producing a book. The distinction becomes less obvious as the assistance grows more extensive. A developmental editor may identify structural problems, recommend changing a character's motivations, or suggest moving an important revelation to another part of the story. Those recommendations can substantially alter the finished manuscript, even when the author performs the revisions personally.
A line editor can also become deeply involved in the prose. They may restructure sentences, remove passages, and propose alternative language that changes how a character or event is presented. More extensive intervention, sometimes described as book doctoring, may involve rewriting entire chapters or reconstructing significant portions of the narrative. Another person can contribute original expression while working on a manuscript that will ultimately carry the author's name. Ghostwriting extends that division further. Someone commissioning a book might provide a detailed outline, a partially completed manuscript, or little more than a premise. The ghostwriter can develop the remaining plot, construct scenes, write dialogue, and produce the complete manuscript. Some clients participate throughout the writing process, while others primarily review the material as it is delivered.
The Authors Guild maintains a marketplace for professional ghostwriters, and contracts can establish compensation, ownership, confidentiality, and attribution. A ghostwriter may receive prominent credit, appear in the acknowledgments, or agree to remain anonymous. The client may have created the underlying fictional property, financed the work, and assumed the costs of publication. The ghostwriter receives compensation for their contribution under the negotiated agreement. These arrangements have existed long enough that publishing has developed ways to manage them without requiring the person whose name appears on the cover to have personally written every sentence. The commercial relationship can be perfectly clear to the people involved, even when readers have little understanding of the division of labor.
When I see a novel carrying someone's name, my natural assumption is that the person wrote it. I understand that editing was probably involved, but I wouldn't automatically assume that someone else created entire chapters or supplied most of the finished prose. Yet commercial publishing has accommodated both possibilities for generations, including continuing series produced through ghostwriters and books substantially revised by other professionals. In Does Writing Make You a Storyteller?, I examined the difference between developing a story and composing the words used to tell it. A credited author may have originated the characters, relationships, and narrative direction, while other people performed substantial portions of the writing. The byline identifies the person publicly associated with the book, although it doesn't necessarily describe the work performed by everyone who contributed to its creation.
Those contributions can also be purchased, and their cost affects which authors can obtain them. An author who can afford sustained professional assistance has access to a different process for developing and revising a manuscript than someone who must work primarily alone. That difference becomes particularly apparent when a book needs more than ordinary copyediting.
The Cost of Professional Assistance
Imagine a writer who has completed an 80,000-word novel. They understand their characters, have developed the central conflict, and know how the story ends. The opening works reasonably well, and the final chapters resolve the major events. Somewhere in the middle, however, the manuscript loses direction. The protagonist spends several chapters reacting to events without making meaningful decisions, relationships remain underdeveloped, and scenes that seemed necessary during outlining contribute little to the finished story. The writer recognizes that something isn't working but cannot identify where the problem begins, and every attempt to repair one chapter seems to introduce problems elsewhere.
An experienced developmental editor can examine the manuscript as a complete work, tracing the relationships between scenes, characters, and narrative decisions. They may recognize that a revelation occurs too early, that a character's motivations change without sufficient development, or that several chapters repeat the same conflict. The editor can explain why those problems have developed and consider how proposed revisions might affect the remaining story. The value of professional experience extends beyond identifying weaknesses. An editor can develop an understanding of the author's intentions, recognize when an unconventional choice is deliberate, and challenge decisions that undermine the manuscript. They can discuss why an earlier recommendation was rejected, reconsider their assessment after reading a revision, and work through disagreements with the writer.
The author may still have considerable work to perform. They might remove chapters, rewrite dialogue, reconsider a character's motivations, or change the order in which events occur. Editorial criticism can influence the finished novel substantially while leaving the author responsible for writing the revised material. The assistance can also become expensive. According to the Editorial Freelancers Association's 2026 rate survey, developmental editing for an 80,000-word fiction manuscript would cost approximately $2,400 to $2,800 using its reported median per-word rate range. Full-length fiction ghostwriting would cost approximately $20,000 to $40,000 at the corresponding rates. These are illustrative figures rather than standard prices, and individual professionals may charge considerably more or less.
Less extensive professional services can be less expensive, and critique partners, writing groups, and beta readers can provide valuable feedback. Traditionally published authors may also receive editorial support through their publishers. But sustained professional assistance, particularly when a manuscript requires extensive restructuring or several rounds of revision, can represent a substantial investment. An established author with a publishing agreement or a commercially successful series may treat that expense as part of producing another book. Someone writing their first novel without a publisher has to consider whether spending thousands of dollars makes sense for a manuscript that may never generate meaningful income. Recognizing that an experienced editor could improve the work doesn't make the service affordable.
A writer unable to purchase professional assistance may spend months attempting to identify structural weaknesses alone. They can make progress through practice, feedback from other writers, and repeated revision, although they may never receive the kind of sustained examination that helps expose problems they have become too familiar with to recognize. A language model provides another option. The writer can supply the problematic chapters, explain what they are trying to accomplish, and ask the model to examine the protagonist's decisions or identify scenes that aren't advancing the conflict. They can request alternative structures, question the suggestions, and compare possible revisions. Some recommendations may reveal problems the author hadn't recognized, while others may be unhelpful or misunderstand the story entirely.
These activities overlap with work performed during developmental editing. The model may not provide criticism of comparable quality, and its limitations become apparent when working through a long manuscript. It can misinterpret narrative intent, invent connections unsupported by the text, or suggest conventional plot developments that don't fit the author's intentions. Its ability to maintain continuity across a novel also depends on the model's context capabilities and how the writer supplies the material. An editor who has carefully read the manuscript and discussed it with the author can develop an understanding of the work through sustained attention and experience. An LLM can analyze many of the same problems, but the reliability of its analysis depends on the system, the information it receives, and the writer's ability to evaluate its responses. Fluent criticism can be persuasive without being particularly good, and inexperienced authors may struggle to recognize the difference.
For someone who can afford professional editing, those limitations may justify hiring an experienced editor. Someone who cannot afford the service faces another calculation. They may obtain imperfect assistance through AI or continue without that additional source of criticism. The opportunity to question recommendations, explore alternatives, and repeat the review can be useful even when the writer rejects most of the suggestions. They still need to understand the manuscript well enough to recognize what belongs in it. AI can reduce the cost of obtaining criticism without eliminating the judgment required to use that criticism effectively.
From Editing to Writing
The amount of creative work delegated changes as assistance moves beyond criticism. Suppose the author asks an LLM to suggest an alternative structure for a chapter that isn't working. The model might recommend introducing a betrayal, revealing a concealed identity, or escalating an external threat. Any of those changes could increase tension, although they might also turn a story concerned with a character's gradual loss of confidence into a conventional thriller the author never intended to write. A model can produce a coherent solution to the wrong problem. An experienced editor can make the same mistake, particularly when commercial expectations influence their interpretation of the manuscript. A sustained working relationship may help the editor recognize the author's intentions, while the writer using AI may have to explain those intentions repeatedly and reject suggestions that conflict with them.
Research offers some reason to examine the effect of AI suggestions on originality. In a 2024 study published in Science Advances, Anil Doshi and Oliver Hauser studied participants writing short fiction with different levels of access to AI-generated story ideas. Readers evaluated the assisted stories more favorably, with particularly pronounced benefits among participants who initially scored lower on a creativity measure. The researchers also found that stories produced with AI assistance became more similar to one another. The experiment concerned short fiction and idea generation rather than developmental editing of complete novels. Its findings nevertheless illustrate a potential tradeoff: suggestions that help an individual writer produce a more favorably evaluated story may also encourage familiar narrative choices across many writers. The author has to decide whether a recommendation improves the particular story they are trying to tell, especially when a proposed change creates immediate tension while damaging character development or undermining something distinctive about the manuscript.
The same responsibility exists when the model begins generating prose. An author might write a chapter and ask AI to identify awkward passages, request alternative wording for a conversation, or have the system produce an entirely new version of a scene. They might eventually provide a detailed outline and ask for a complete chapter. These requests delegate different amounts of work. Correcting grammar is unlike generating new dialogue, and producing a complete chapter involves more direct composition than evaluating one the author has already written. Human editors, book doctors, and ghostwriters perform similarly varied services, with differing levels of responsibility for the finished manuscript.
Someone who asks AI to identify problems and personally rewrites the affected material has made a different contribution from someone who accepts a generated chapter with few changes. A ghostwriting client who provides detailed outlines and substantially revises every draft has also participated differently from a client who approves a completed manuscript after supplying little more than the initial premise. The assistance tells us something about how the work was produced, while the amount of creative labor delegated depends on the particulars of the arrangement. Publishing already recognizes those differences when describing human services. A copyeditor, developmental editor, book doctor, and ghostwriter undertake different responsibilities, and the same distinctions remain useful when software performs some of their tasks.
What Publishing Already Accepts
The industry's response to AI becomes interesting when placed beside its established use of ghostwriters and editors. The Authors Guild's Human Authored certification identifies books whose text meets its definition of human-produced writing, allowing only a very small amount of AI-generated or AI-modified material. The certification is concerned with the origin of the text, a narrower question than determining how much of the manuscript was personally composed by its most prominently credited author.
The Guild also maintains a professional marketplace for ghostwriters. Its certification definition refers to text written by one or more humans, although the published registration rules do not clearly resolve every arrangement involving substantial anonymous ghostwriting. It would therefore be inaccurate to claim that every ghostwritten book automatically qualifies. What the comparison demonstrates is that a book can contain substantial contributions from human collaborators while the credited author retains public recognition for the finished work. A designation concerning whether humans produced the prose cannot, by itself, establish which human wrote it.
Amazon Kindle Direct Publishing makes a different distinction. Its rules separate AI-generated material from AI-assisted content. Authors who produce the material themselves and use AI for editing, refinement, or brainstorming generally aren't required to disclose that assistance to Amazon. Content generated by AI must be disclosed to the platform, even if a human subsequently edits it substantially. Amazon's policy recognizes a practical difference between using software to improve writing the author produced and using software to generate the content being published.
A publishing platform or literary organization may adopt its own standards for the types of assistance it permits or recognizes. Those rules can identify how a manuscript was produced, but they don't necessarily measure the named author's original creative contribution. An author can employ another human to write a substantial portion of a book and receive public credit under an accepted contractual arrangement. Another author can personally write and revise a manuscript with extensive AI criticism. Depending on the rules being applied, those works may receive different classifications even when the second author personally performed more of the writing. This becomes particularly consequential when the accepted forms of human assistance require money that some writers do not have.
Who Can Afford the Help?
An author with sufficient financial resources can hire a developmental editor, commission several rounds of criticism, arrange substantial rewriting, or employ a ghostwriter to produce a manuscript. The professionals provide experience and labor that deserve compensation, and the assistance can improve the finished book considerably. Someone with limited financial resources may have none of those options, particularly when the manuscript needs extensive work. They may have the same determination to complete a novel and the same willingness to revise it, but obtaining another person's sustained attention and expertise can be prohibitively expensive.
AI makes portions of that assistance available at a much lower direct cost. It doesn't eliminate the value of professional experience or guarantee a better manuscript. It gives writers another way to examine problems, explore alternatives, and revise work they might otherwise have struggled through alone. There is also an economic distinction between replacing professional services and making them accessible to someone who couldn't purchase them. An author who previously paid for developmental editing may decide that AI can handle enough of the work to reduce future editorial expenses. That decision could reduce paid assignments for professionals, while the effect on writers who were never able to afford those services is different.
Someone who cannot purchase a $2,400 developmental assessment isn't necessarily choosing AI over an editor. Hiring the editor may never have been realistic, and their use of AI doesn't automatically represent a professional assignment that would otherwise have existed. Writers with greater resources can also use AI alongside professional editing, obtaining feedback from software before commissioning more specialized assistance. The technology reduces some financial barriers without eliminating the advantages money provides throughout publishing.
What interests me is the difference in how we may judge the authors using those options. Someone who can afford substantial human assistance can employ established publishing services, potentially delegating extensive portions of the creative work. Someone who personally writes and revises a manuscript using inexpensive AI assistance may have contributed considerably more of the finished prose. The cost of the service doesn't establish which person exercised greater creative judgment, and the identity of the assistant doesn't tell us how much work the author delegated.
Consider the writer who has completed a manuscript, knows it needs improvement, and cannot afford the professional services that might help. A language model may provide flawed criticism, suggest inappropriate revisions, or encourage familiar narrative conventions. The author still has to decide what belongs in the story, reject changes that undermine it, and perform the work required to produce the manuscript they intended. An author with sufficient resources can pay someone to provide that criticism, rewrite substantial portions of the work, or compose the entire book. Publishing has accommodated those arrangements for generations, including situations where the person receiving public credit wrote relatively little of the finished prose.
I have difficulty accepting the assumption that extensive purchased assistance is an ordinary part of publishing while less extensive assistance through AI makes an author's contribution suspect. The two methods aren't identical, and their quality can differ considerably, but we're familiar with the underlying practice of obtaining help to create and improve a book. AI has made some of that help affordable to people who would otherwise have gone without it. I don't see why those writers should face a greater expectation of personal creative labor simply because they couldn't afford to hire another person.