Sunny Greenhill – How to Write a Books with ChatGPT (страница 3)
●
build the scene;
●
change perspective;
●
achieve a coherent voice;
●
do the editing.
If you do not do that, your control starts resembling operation of a machine rather than creation of a work.
Practice: how to make your contribution obvious, even if you use AI a lot
Imagine that a year from now you will want to:
●
sell translation rights;
●
sign a contract with a publisher;
●
defend yourself against a claim that "this is not your text";
●
or simply prove your authorship calmly.
Not stories, but traces of your work.
I would advise a beginning author to develop a habit that at first seems boring, but later becomes a lifesaver: keep the process as if one day you might have to show "how this was made."
Not publicly. For yourself.
For example:
●
a book plan (at least in draft form);
●
a chapter structure;
●
character notes;
●
a scene draft;
●
your reworking;
●
the final version.
This is not bureaucracy. It is "provability."
And what matters is this: you do not have to keep everything. It is enough that you can show the logic of your decisions.
A separate knot: what if AI generates, and you only select
There is a common question: "If I do not rewrite every line, but simply select the best – is that a contribution?"
Yes, it can be a contribution, and a fairly serious one. Editing is creative work. An editor is a profession. Curation of material can also be a creative act.
But there is a subtle point here: the selection must not be "everything in a row" and not "anything, as long as something exists." It has to shape your intention. And ideally, it should be accompanied by reworking of the voice. Then it becomes very convincing.
If the selection is mechanical – "I take the first answer that is more or less okay" – that is closer to operation than to authorship.
What you should take away from section 1.2
In legal reality, the dispute is usually not about "is AI the author or not," but about how visible and provable your authorship in the book is.
Different countries and systems may treat minimal human contribution differently, but all of them respond well to a situation where the human being:
●
controls the structure;
●
makes creative decisions;
●
assembles the text into a coherent work;
●
and can prove that it was their work, rather than an exported result.
And that is already a practical foundation: if you build your process so that your contribution is not cosmetic, but substantive, you close most of the risks before they even appear.
1.3. Your Role and Your Responsibility: Why AI Is Never at Fault, and How That Affects Rights, Risks, and Publication
There is a point that many beginning authors do not want to hear at first. It sounds unpleasant because it takes away comfort. But it is exactly what makes your position mature and stable:
That means if something goes wrong in your book – no matter on whose suggestion, through whose "generation," in whose style… it will not be the service, the model, or the "algorithm" that answers for it. You will answer for it, because your name is on the cover, you clicked "publish," you received the money, and you present the text to the world as your work.
I want you to take this not as a threat, but as a point of support. Because responsibility is not punishment. It is the anchor that ultimately makes you the author: you are not merely playing with words, you are carrying the consequences.
Why this is especially important in the question of "co-authorship with AI"
When a person says, "we wrote it together," they are often unconsciously trying to divide responsibility in the same way they divide the work. As if, if it is written poorly – "AI messed up." If there is an error – "AI made it up." If there is some questionable passage – "that is just how it was generated."
But in reality, that is not how it works. And the sooner you stop thinking in that logic, the fewer unpleasant surprises there will be later.
The law thinks simply:
AI, unlike a human being, cannot: