Why AI Can’t Write the Book You Actually Need
- Jun 24
- 6 min read
Taryn Lee Johnston is a UK publisher, writing partner, and integrated media consultant specialising in helping high-achieving professionals write and publish books that build authority. She is the founder of FCM Publishing, Chronos Publishing, and Taryn Lee Johnston Integrated Media Consultancy.
Artificial intelligence may be changing how books are written, but for authors who want their work to carry authority, personality, and lasting impact, speed is not always the advantage it appears to be. This article explores what is lost when the writing itself is handed over to a machine, and why a book's greatest value still lies in the distinctive human voice behind it.
What you lose when you let the machine do the work
If you are thinking about writing a book, the question of whether to use AI to do it has almost certainly come up. The tools are visible everywhere, your peers are likely already experimenting with them, and the appeal of cutting twelve months of solo writing down to a few weekends is obvious for anyone running a business while also trying to produce something meaningful.
There is a place for AI in writing a book, and pretending otherwise would be the wrong response. Transcription, summarising long documents, generating outlines you might play with, inspiration for those dry moments, drafting marketing copy, and formatting text, all of these are reasonable uses where the time savings are real, and the loss of quality is small.
What AI is incapable of doing, and what most business owners writing a book do not realise until the work is finished, is producing the actual book that will represent them in the world for the next decade.
There are three reasons this matters more than the appeal suggests:
The first reason: The industry can spot it a mile away
Publishers, editors, agents, reviewers, and anyone else who reads professionally for a living can all recognise AI-generated writing within a paragraph or two, because the patterns are too consistent to disguise and the cadence betrays itself in ways that human writers rarely produce. The arguments tend toward a smoothed-over quality with no rough edges, the structure of paragraphs follows shapes that an experienced reader registers without needing to name, and certain phrases recur at a frequency that flags the source immediately to anyone who has been paying attention.
This matters because the people you most want to be impressed by your book are the people who will spot the machine first. The journalist who might quote you, the conference organiser who might invite you to speak, the senior figure in your industry who might mention your book in a meeting, the reviewer who might write about it, not to mention the publisher you hope will represent you, all of them have been reading enough AI output to recognise it on sight. If any of these people open your book and identify it as AI-generated, the entire purpose of writing the book is gone.
The harder part is that the recognition is rarely something they will tell you about, I would, but I may be in the minority. They simply put the book down and move on, and you will not get a phone call explaining why they did not engage with your work. You will just notice, over the months that follow, that the book is not doing what you hoped it would, and you will have no clear way of knowing why.
The second reason: Many ordinary readers can spot it too
The assumption that AI-generated text only registers as AI to industry professionals is becoming less true every month, because readers are exposed to enough machine-written content now that they recognize the pattern at a subconscious level, even when they cannot put a name to what they are noticing. They will read a few pages, feel that something is slightly off about the voice, and slowly stop reading. What they tell themselves afterward is usually some version of "this book was not really for me," when what is actually happening is that they are picking up on the absence of a human behind the words and losing interest accordingly.
By the time a few people in your network have read enough of your book to feel that way, the word has already traveled that the book is not really worth recommending, and again, you will not be told. You will simply notice that the book is failing to generate the conversations you expected it to spark.
The other thing happening with readers right now is that the volume of AI-generated content across every medium has made authentic voices stand out more rather than less. People are actively craving writing that sounds like a real person thought it through, and a book that does not sound like a real person becomes background noise from the first page.
The third reason: You will lose the whole point of the book
The reason for writing a book at a senior level is to put your distinctive thinking in front of readers in your own voice, so that the book becomes a representation of how you see the world, how you talk, how you reason, and what you actually believe. That is the asset, and it is what justifies the time and the cost of writing it properly.
The moment you let AI generate the prose, you lose all of that, because AI smooths everything toward an average and removes the phrasings that are characteristic of you, replaces them with phrasings that are characteristic of nothing, and produces a book that could have been written by anyone or no one. The book becomes a thing that exists rather than a thing that represents you, which is exactly the opposite of what you set out to achieve.
These are the trade business owners who actually care about the book, and they do not see when they encounter AI book writing tools. This argument only matters if you actually care what the book does once it exists. For the business owner who only needs "author" in their bio, the AI tools serve the purpose perfectly well. For the business owner who wants the book to represent them and open the doors that are currently closed to them, what AI removes is exactly what the book was for.
What this means in practice
If you are thinking about writing a book and AI is being suggested as the solution, the practical advice is straightforward. What you should avoid is using AI to generate the actual prose of the book, because the drafting of the sentences is where your voice lives, and the moment you delegate that work, the book stops being yours.
The tools that promise to write the book for you are not selling you a shortcut, they are selling you a different product that happens to look like the one you wanted.
If the prospect of writing a book at the standard you actually need feels too large to do alone, the answer is to work with a partner who can do the structural and developmental work alongside you, because that is the work AI is incapable of doing anyway. A serious partner who understands how senior books actually function is the alternative that AI marketing is hoping you will not consider.
The book that speaks for you for the next decade, at least, is worth the time it takes to write properly. The authors I work with recognise this sooner rather than later, which is why the book they hold in their hand truly represents them, and why they are proud to put their name on it.
Read more from Taryn Lee Johnston
Taryn Lee Johnston, Publisher, Book Writing Partner, Strategist
Taryn Lee Johnston is an independent publisher, UK writing partner, and thought leader on artificial intelligence in publishing. Founder of FCM Publishing and Chronos Publishing, she has worked with clients including Tony Robinson OBE and Bob Champion CBE. Named Independent Publisher of the Year 2024/25/26, her authors have achieved six consecutive Business Books of the Year shortlistings. She is also a TEDx speaker and a former University of Lincoln lecturer.
