[Journey] Creating Good Books With AI

[MasonBookwright]

So I finally signed up to Blackhatworld after being a little scared of the place.

But I have gotten over myself and now I am here.

As I mentioned in my intro I have previously made money online through Adsense and affiliate marketing until the big G changed direction and tanked all my sites.

Then there’s my current business major POD mugs with AI art. It’s doing okay but even in POD there are unexpected costs and pitfalls.

I have decided to experiment with writing non fiction books with AI. I have previously done some content with it and know a little bit about the pitfalls and some tips and tricks to get better output.

My ultimate aim is not to create AI slop but actually useful material that can stand the test of time.

I have chosen to mainly use Claude for the books as I like the projects and being able to keep things together but I am not afraid to try others.

So here’s the plan:

Decide on a topic
Create a book outline
Create an author and author voice
Write the book
Redraft
Edit

My estimated costs are £20 for Claude but open to spend more as the journey goes on.

Goals:

Prompt
Publish
Profit

I am going to start with a random non fiction topic probably in self help or wellness.

When the process is complete I will look at market research for choosing the books to launch. For the moment I want to sort out the book writing process first.

The first thing I have seen being taught is to outline the book. So that’s my first job.

Here’s the prompt I will start with:

[You are a best selling author specialising in books on [niche] . You have been tasked with outlining a best selling book on [niche]

I am going to pop this into chatgpt Claude grok and Gemini and see what comes back.

Following this I am going to pop across to Amazon and do a search and see how the book chapters of actual books compare to what AI has spit out.

Anything I am missing at this stage?

Any questions please just ask.

 

[Panther28]

The problem is there are 10's of thousands of people publishing crappy ai books like this. Ai, can't write witht he creativity of a human brain. If you think it can you are either very uncreative, or english isn't your first language?

Write something in an area that you are interested in. Pick something that you can have a natural advantage with, and know about extensively that is niche, and write it in your perspective, as fast as you can type on the keyboard, and do that for 30 days. You will then have a book that will be better than an AI title, and maybe if it doesn't sell much. You'll appreciate the difficult journey of authors, instead of this make believe world, where we all write 100 books, and become millionaires! lol

Best of luck

 

[wolfofdevstreet]

panther is right, but at the same time it is not impossible to do it with ai and still get sales, you need to know how to prompt it tho, and it wont be like people think, pop 1 prompt get a full e book and become rich, no you will need to prompt it many many times until its actually an ok book, it might be faster than sitting there and writing a long book yes, but its not easier, and you have to sit and check everything, and you have to not only see it from your own perspective, you have to see it from the target audience perspective, but if its only 20 bucks, hell give it a try and see what happens, what do you got to lose.

 

[MasonBookwright]

The problem is there are 10's of thousands of people publishing crappy ai books like this. Ai, can't write witht he creativity of a human brain. If you think it can you are either very uncreative, or english isn't your first language?

Write something in an area that you are interested in. Pick something that you can have a natural advantage with, and know about extensively that is niche, and write it in your perspective, as fast as you can type on the keyboard, and do that for 30 days. You will then have a book that will be better than an AI title, and maybe if it doesn't sell much. You'll appreciate the difficult journey of authors, instead of this make believe world, where we all write 100 books, and become millionaires! lol

Best of luck

thank you I appreciate your response and I largely agree with what you are saying - there is a ton of slop out there.

I can well believe 99% of people using AI to publish are slapping a generic prompt into ChatGPT, and dumping an unformatted PDF straight onto Amazon expecting to make millions overnight.

But just because the market is flooded with bad AI books doesn’t mean you shouldn’t use AI. It is like saying you shouldn't build a niche website because there are tens of thousands of crappy WordPress blogs out there.

But that’s exactly why I started this journey.

I’m not doing a "one-click" spam method.

I want to see what AI is actually capable of. Through trial and error can I get AI to write a good.

If I can’t I won’t publish anything.

 

[MasonBookwright]

panther is right, but at the same time it is not impossible to do it with ai and still get sales, you need to know how to prompt it tho, and it wont be like people think, pop 1 prompt get a full e book and become rich, no you will need to prompt it many many times until its actually an ok book, it might be faster than sitting there and writing a long book yes, but its not easier, and you have to sit and check everything, and you have to not only see it from your own perspective, you have to see it from the target audience perspective, but if its only 20 bucks, hell give it a try and see what happens, what do you got to lose.

I agree it’s not going to be a one and done or a two and done prompt thing.

I remember using AI to create a skyscraper post way back when and even in that it struggled.

By the end it was getting names and places wrong.

Then you’ve the obvious tells or triplets and set phrases or uses let alone the fact it will randomly make something up.

AI will have developed since but it will be interesting to see if it’s any better.

 

[S1810303350]

So I finally signed up to Blackhatworld after being a little scared of the place.

But I have gotten over myself and now I am here.

As I mentioned in my intro I have previously made money online through Adsense and affiliate marketing until the big G changed direction and tanked all my sites.

Then there’s my current business major POD mugs with AI art. It’s doing okay but even in POD there are unexpected costs and pitfalls.

I have decided to experiment with writing non fiction books with AI. I have previously done some content with it and know a little bit about the pitfalls and some tips and tricks to get better output.

My ultimate aim is not to create AI slop but actually useful material that can stand the test of time.

I have chosen to mainly use Claude for the books as I like the projects and being able to keep things together but I am not afraid to try others.

So here’s the plan:

Decide on a topic
Create a book outline
Create an author and author voice
Write the book
Redraft
Edit

My estimated costs are £20 for Claude but open to spend more as the journey goes on.

Goals:

Prompt
Publish
Profit

I am going to start with a random non fiction topic probably in self help or wellness.

When the process is complete I will look at market research for choosing the books to launch. For the moment I want to sort out the book writing process first.

The first thing I have seen being taught is to outline the book. So that’s my first job.

Here’s the prompt I will start with:

[You are a best selling author specialising in books on [niche] . You have been tasked with outlining a best selling book on [niche]

I am going to pop this into chatgpt Claude grok and Gemini and see what comes back.

Following this I am going to pop across to Amazon and do a search and see how the book chapters of actual books compare to what AI has spit out.

Anything I am missing at this stage?

Any questions please just ask.

Market Research First, Not Last: It's highly recommended to do market research before you outline or write the book, rather than waiting until the process is complete. You need to verify if there is an actual buying audience for that specific self-help/wellness sub-niche on Amazon.

 

[MasonBookwright]

Market Research First, Not Last: It's highly recommended to do market research before you outline or write the book, rather than waiting until the process is complete. You need to verify if there is an actual buying audience for that specific self-help/wellness sub-niche on Amazon.

I won’t be selling the first few books. It’s to experiment and test concepts. Once that’s done then yes I agree writing to market will be important.

 

[MasonBookwright]

[Experiment 1]

Can AI create a comprehensive book outline that covers the subject matter?

How this test will be run?

I will test three prompts for the outline

Prompt One: Outline a book on …..

Prompt Two: You are a best selling author specialising in books on [niche] . You have been tasked with outlining a best selling book on [niche]

Prompt Three:
You are an expert author, psychologist, and philosopher, equal in knowledge, background, and philosophical approach to [Ichiro Kishimi]. Your expertise is deeply rooted in [Adlerian psychology], focusing on concepts like [the separation of tasks, interpersonal relationships as the source of all problems, and the courage to be disliked].

You have been tasked with creating a comprehensive chapter-by-chapter book outline on the topic of: [People Pleasing].

The target reader for this book is [someone who constantly sacrifices their own needs for others, feels trapped by societal expectations, and feels immense guilt when saying no].

Please generate the book outline according to the following constraints:
1. Tone & Style: The tone should be philosophical yet highly accessible, compassionate but firm, and deeply insightful. [Optional: Deliver the outline reflecting a dialogue structure between a frustrated individual and a wise guide.]
2. Structure: Break the book down into logical parts (e.g., 5 major parts), with 3-4 chapters per part.
3. Chapter Detail: For each chapter, provide:
- A compelling, philosophical Title.
- The Core Psychological Shift (what the reader realizes by the end of the chapter).
- 3-4 Bullet Points detailing the specific concepts, psychological theories, or shifts in perspective discussed in that chapter.
4. Core Theme: Ensure the entire outline weaves back to the central idea that [true freedom comes from the willingness to be disliked and taking responsibility for one's own tasks].

Begin by giving a brief 2-sentence summary of the book's overarching thesis, and then dive straight into the outline.

It will be interesting to see what they bring back and what similarities and differences there are between the prompts.

To benchmark the responses I will search the best selling books in this niche and collate the information in each book. I aim to do this both manually and with Notebook LM.

I will then compare the information in the human books to the AI outline checking for

Coverage of benchmark topics	35%
Logical flow	20%
Missing important concepts	15%
Redundancy	10%
Practical usefulness	10%
Originality/added value

Which prompt scores highest will be used. If everything falls below an 80% mark then the prompts will be revisited.

Should take about a day to do this around my other commitments.

Part of this feels like my old seo days going through the SERPs looking at H1’s

 

[MasonBookwright]

Experiment 1 Results.

I tested ChatGpt and Claude in the end on all three prompts. Wanted to Grok but it was always too busy to render a response on the free plan.

Prompt 1

ChatGPT failed to meet the benchmark producing a generic outline resembling more of a blog post than a useful book

Claude was slightly more creative and pulled together more common features but lacked anything extra.

In SEO terms this was like using a tool to extract headings from articles and then merging them together and listing them in a arbitrary list of importance.

Prompt 2

Significantly better result from ChatGPT with a more extra dynamics added. On this test it benchmarked closer on coverage of topics but agin missed important concepts.

Some of the suggested chapters would have been better folded into one chapter

Claude seemed strangely worse with this prompt than before. It again provided nothing in terms of originality and also missed some quite large chunks of needed info. It was like it had decided what the target audience was.

Prompt 3

Both did better here but both suffered from the same lack of logical flow. I think the naming of someone provided too much direction. They both mirrored quite closely the authors original work but didn’t seem to expand the topic any further.

I upped the level on Claude and it doubled down on this issue. As a further thought I also ran the second prompt on a higher level and got much better results.

Conclusion

A single prompt fails to adequately create a profession book outline that matches with the best selling books in this niche. A multipart prompt or alternative AI may fair better.

An interesting experiment would be to place ebooks of the best sellers into Notebook and see if it could handle the request. I feel that although Notebook would extract the information providing an internal logic to the book may be a struggle.

Next step

Claude Opus Max handled the requests best and benched mark top of the comparison chart.

I will use this with the second prompt to outline a book. At this stage I won’t be manually editing or going to a multiprompt system.

Next step will be voice creation and testing chapter 1 with different voices to see how the AI writes.

 

[mysticparadise]

It's not only a question of how many AI books are out there, but also competition in general. By now there are probably hundreds of quality books written by actual human writers on every topic. In fact, many writers hire ghost writers from 3rd world country that are happy to write an entire book for a small compensation because in their country it's a lot of money, even for a month of writing. And they happen to have some pretty good English. Good enough to be a good first draft that a native speaker can edit later.

So you would need to find niches where the competition is very low, since these other books already have lots of reviews, and people would prefer to buy their books over yours.

Plus, I think if you are certain you want to write the book with AI, it has to be a long dialogue. On their own they'll never write something good from A to Z, no matter the prompt. So you have to work on each page, or even each paragraph separately and check it against your own understanding of how things should look.

 

[MasonBookwright]

It's not only a question of how many AI books are out there, but also competition in general. By now there are probably hundreds of quality books written by actual human writers on every topic. In fact, many writers hire ghost writers from 3rd world country that are happy to write an entire book for a small compensation because in their country it's a lot of money, even for a month of writing. And they happen to have some pretty good English. Good enough to be a good first draft that a native speaker can edit later.

So you would need to find niches where the competition is very low, since these other books already have lots of reviews, and people would prefer to buy their books over yours.

Plus, I think if you are certain you want to write the book with AI, it has to be a long dialogue. On their own they'll never write something good from A to Z, no matter the prompt. So you have to work on each page, or even each paragraph separately and check it against your own understanding of how things should look.

I get what you are saying but I believe there will never be enough books.

One of the beauties of book buyers is they buy books.

Dieters can have dozens of books all saying the same thing.

Conspiracy theorists all look for the next revelation.

Sale speople consume all sorts of selling strategies psychology and play books.

I see competition at the market being validated. That said there are underserved and developing niches that are worth looking at in time.

I agree with your other point that one prompt won’t create a book of any substance. I need to do this first for myself so that I understand the process and can see the final output. What does it look like why does it do that.

It’s only from there I can look at developing a more detailed better work flow.

As an example today I asked AI to “write the first chapter” and it did this

“
Most people who struggle with people-pleasing don’t recognize it as a problem at first. In fact, it often looks like a strength. Being “easygoing,” “helpful,” “kind,” or “low maintenance” is socially rewarded. You might even be praised for being the person who never causes trouble, who always says yes, who can be counted on.

But people-pleasing isn’t just kindness.

It’s kindness that comes at your expense—and often without your real consent.

At its core, people-pleasing is a pattern of prioritizing other people’s comfort, approval, or emotional stability over your own needs, preferences, and boundaries. It’s not simply doing nice things for others. It’s doing them automatically, anxiously, or at the cost of self-abandonment.

That distinction matters.

Because healthy kindness is a choice. People-pleasing often feels like an obligation.

The Invisible Shift from Choice to Compulsion

Everyone does nice things for others sometimes. You help a friend move, you listen after a hard day, you compromise on dinner plans. That’s normal, reciprocal social behavior.

People-pleasing begins when those choices stop feeling optional.

Instead of:

• “I want to help, so I will.”

It becomes:

• “I should help, otherwise they’ll be upset.”
• “I can’t say no.”
• “It’s easier if I just agree.”

Over time, the internal question shifts. You stop asking What do I want? and start asking What will keep things smooth?

That shift is subtle. And because it happens gradually, many people don’t notice it until they feel exhausted, resentful, or strangely disconnected from themselves.”

As raw output I think we can all agree it’s junk. Eventually I want to look and see why it’s junk why AI writes like that and ultimately how to make it work.

 

[exiledvai]

i am vbery curious if this actually works..

 

[MasonBookwright]

I have now made 2 books from AI following the same methods as above. As expected both are junk. I am very pleased with that.

I am going to widen my reading and buy some crap obviously AI books off Amazon.

I am reminded of a story about people who work in counterfeit currency. The story goes they never study counterfeit notes but spend all their time becoming experts in the real notes.

By mastering every detail of the real thing they can spot a fake with ease. It’s probably a fictional story but it has stuck with me for years.

I hope it works the other way around.

As an aside I have noticed there is a dip in the amount of YouTube gurus posting on KDP at the moment. Wonder if demand is down in the age of AI for their courses.

 

[MasonBookwright]

My last thought about the gurus sent me down a bit of a rabbit warren this weekend. I have been listening to hours of tutorials hacks and insider secrets from people who all made a gazillion dollars from publishing AI books.

Everything they teach seems to be

Find a low competition high demand niche

Ask chat to outline the book chapters

Ask chat to provide a break down of the chapters

Ask it to write

Ask it to edit

Publish

Is that what other people have seen?

It moves me a little closer to understanding why these books suck. I think we have two problems going on. There is an inherent problem with AI writing and the input from users aren’t sufficient.

The first one is more problematic but the second one should be trial and error.

 

[Onemarket]

That's a funny idea, buddy. The publishers definitely won't let you get away with it.

 

[MasonBookwright]

That's a funny idea, buddy. The publishers definitely won't let you get away with it.

What do you think they won’t like?

 

[Stavro]

What do you think they won’t like?

If a publisher even suspects you've used AI, they'll remove the listing until you can prove it's not AI.

 

[MasonBookwright]

If a publisher even suspects you've used AI, they'll remove the listing until you can prove it's not AI.

I will be self publishing through POD and KDP.

Amazon care about the end user more than they do if it’s AI hence the mission to not produce slop..

I think it will only be a matter of time before publishing houses embrace AI. Mills and Boon and James Patterson already have writers bibles that they give to ghostwriters.

 

[Stavro]

I will be self publishing through POD and KDP.

If they get any complaints form customers about the book being AI, they'll take a remove first, ask questions later approach.

 

[MasonBookwright]

If they get any complaints form customers about the book being AI, they'll take a remove first, ask questions later approach.

I get you and that’s a very real concern.

To some extent that’s the point of this. Is it possible to make good books with AI. where readers don’t complain or feel cheated.

Of the books I have been reading it’s obvious why people would complain. They are essentially information regurgitation systems. They lack that quality that makes a book a book and not just what GPT spits out.

I think all authors are going to have to face the question why would a reader read your book and not AI the answer.