Sunny Greenhill – How to Write a Books with ChatGPT (страница 10)

Because the moment you write "Author: me and ChatGPT," you create the shadow of a third party. And even if everyone understands that AI is not a subject, the carelessness itself looks like weak legal culture. And the market reacts poorly to weak legal culture.

And this is not a matter of taste. It is a matter of whether you want your book to move through deals easily or whether you want someone to ask every time: "and what did you mean by that?"

The second mistake is to publish large amounts of generated text "as is"

A beginner has a dangerous joy: AI gives a lot of text quickly. And it feels like a gift. But this is exactly where people often undermine their own position.

When you publish a book that is essentially a "first answer" simply assembled into a file, you create two problems at once.

The first is artistic: the text will sound even, but empty.

The second is legal: your contribution will be hard to show.

If you want to use AI a lot – that is fine. But publishing "raw material" without authorial editing is like releasing a movie from unprocessed takes. Yes, technically it is material. But the work appears later.

The third mistake is to use AI as a substitute for understanding

This is especially important for nonfiction and guides, but it also appears in fiction, when the author starts inserting "smart explanations."

AI can write a convincing paragraph about anything. And a beginner sometimes leaves it in simply because it "sounds good."

And then what I call "the shame of verification" happens: someone asks a simple question – and the author cannot explain their own text.

Legally, this turns into a risk of liability for false information.

Creatively, it turns into a loss of trust in the authorial voice.

In the market, it turns into a weak reputation.

Remember one simple discipline: if you cannot retell the idea in your own words, it should not be in the book.

The fourth mistake is to write "about real people and companies" without caution

AI easily makes text convincing and sharp. It does not feel the boundaries of reputational risk.

A beginner writes: "this company is made up of fraudsters," or "this person did such and such," or "this industry is criminal."

In a work of fiction, this can become "recognizable defamation" if the character is easily read as a real prototype. In nonfiction, it is a direct risk of claims.

AI will not save you in that case and will not take responsibility. You published it – you are responsible.

If you want to write about reality – either hold firmly to the facts, or change the details so that it is artistic rather than "a finger pointed at a specific person."

The fifth mistake is to make "collections" from other people's materials, thinking that AI will clean everything up

There is a temptation that is especially strong for beginning nonfiction authors:

"I am going to gather information from different sources right now, run it through AI, it will rewrite it – and it will be unique."

That is not how it works.

A rewritten text does not always become an original work. It can be rewritten in such a way that the structure, logic, and recognizability remain someone else's. You can "redress" someone else's material and still leave it essentially someone else's.

This is one of the most dangerous zones, because the author sincerely thinks that they "made it their own," while in fact they simply refaced someone else's work.

The correct path is different: not to "rewrite," but to create an authorial structure, an authorial view, and an authorial system of presentation. Then you really are the author, not a compiler.

The sixth mistake is to promise the reader something you are not prepared to answer for

Beginners love loud promises on the cover and in the description: "you are guaranteed to get a result," "precise legal advice," "a reliable way to do such and such."

AI helps you write confidently. But confidence in the text is not equal to a guarantee in reality.

If you write a practical book, especially about money, immigration, laws, or health – any "guaranteed" becomes a potential conflict. And if a reader applied the advice and suffered harm – they have a question for you, not for the algorithm.

Write the way an adult author writes: carefully, precisely, with reservations where they are honestly needed, without "magic guarantees."

The seventh mistake is to keep "nothing" and hope that "it is obvious anyway that it is mine"

This looks like a small thing, but in practice it is exactly the absence of traces that often makes an author weak in negotiations.

If you kept neither the plan, nor the versions, nor the notes – all you have left is the final file. And you may be one hundred percent the author, but it will be harder to prove if someone starts asking questions.

You do not need to turn into an archivist. But leave at least a minimal line of process: concept -> structure -> versions.

The eighth mistake is to think that "if I honestly disclosed AI, then responsibility was divided"

This is very human logic: "I warned them – so I am protected."

In legal logic, that is almost never so.

Acknowledging use of the tool does not remove responsibility for the content. It may increase trust as transparency, but it does not work as a shield.

The shield is quality control. The shield is your decisions. The shield is editing and verification.

Conclusion of Section 1.8

If you gather everything into one phrase, it will be this:

the biggest legal problems with AI are born not from the technology, but from the author's carelessness.

Carelessness is:

to call the tool a co-author in a legally significant place;

to publish raw material without processing;

to leave in the book what you do not understand;

to speak about real people and facts without caution;

to compile someone else's material, thinking that "rewritten means yours";

to promise what you are not prepared to answer for;

not to preserve the trace of the process;

to think that "transparency" replaces responsibility.

If you avoid these mistakes, you are already a head above most beginning "AI authors." And your rights will exist not on paper, but in reality.

1.9. How to Speak Calmly with a Publisher, Editor, and Partners: What to Say About AI So That People Take You Seriously

When you move from the mode of “I write for myself” into the mode of “I publish, sell, and sign,” not only does the scale change… the language changes as well. And here many authors, especially beginning ones, fall into a trap: they begin either to justify themselves or, on the contrary, to flaunt AI, as if that were the main argument for quality.

But in a professional environment, everything is much simpler. People are not interested in your excitement or your fears. They are interested in three things:

who the author and rightsholder are;

whether the text can be used safely (translated, published, sold);

who bears responsibility.

If you know how to answer these questions calmly and clearly, the conversation goes easily. If you start floundering, people see you as a risk. And not because “they are against AI,” but because nobody wants legal surprises.

Let me show you how to speak so that you sound like a mature author.

Let us begin with the main point: your position must be simple