Author, Reader, and AI User. The Problem Is That I Agree with All of Them.

I heard an AI warning at the end of an audiobook. Ever since, my inner gremlin has been holding meetings with no conclusions.

I first heard this wording at the end of an audiobook on Audible, and before you judge me, I should admit that the mere fact I made it all the way to the end credits probably says more about me than it does about the audiobook. Normally, the story ends, people close the app, and peacefully move on with their lives. I, on the other hand, stayed to listen to the copyright information, publisher details, and other bits of information that most people find about as fascinating as the instruction manual for a toaster. I’m not entirely sure why I do this. It’s probably the same manufacturing defect that makes me read shampoo labels, medication leaflets, and footnotes that authors secretly hope nobody will ever notice. What matters is that, while the narrator calmly and professionally continued listing legal information, I heard something that made me look up and rewind.

The voice explained that the book in question could not be used for training artificial intelligence systems. For a few seconds, I was convinced I had missed an important chapter and accidentally wandered into a science fiction novel. But it wasn’t about characters, a technological conspiracy, or some dystopian future. It was simply the copyright section. And of all the places where I expected to find a reflection on the relationship between humans and artificial intelligence, that was probably the last one. I listened again, and that was when my inner gremlin—that character responsible for most of my bad ideas and nearly all of my pointless curiosities—immediately went on high alert. For hundreds of years, authors have wondered whether people were reading their books. In 2026, apparently, they also have to wonder whether robots are reading them. And that was exactly when I realized just how absurdly fascinating this discussion is.

At that moment, my inner gremlin inevitably appeared.

– So, you’re on the authors’ side?

– Yes.

– And you’re on AI’s side?

– Yes.

– That’s not how this works.

– I know.

– Then whose side are you on?

– Yes.

– I hate it when you do that.

– I also hate debates where everyone has good arguments.

As a writer, I completely understand where authors are coming from. A book doesn’t appear out of thin air, nor does it magically materialize between two cups of coffee and a free weekend. Behind it are months or years of work, research, rewrites, personal experiences, doubts, temporary abandonments, and stubborn returns. There are nights when you stare at the same page wondering whether you’ve written something intelligent or simply rearranged the same words in a different order. There are chapters that disappear after months of work and ideas that refuse to cooperate precisely when you need them most. If I spent three years writing a novel and discovered that an algorithm had read it faster than I can finish a bag of pretzels during a Teams meeting, I would probably have a few questions too. A lot of questions. Possibly enough to require a separate appendix and, eventually, a few lawyers.

The problem is that I am probably the last person who should be giving a speech against artificial intelligence. I use AI almost every day. I use it in my profession. I use it for analysis, research, structuring information, and organizing ideas. I use it when I write, and sometimes it helps me spot an angle I missed, identify a logical gap, or articulate my own arguments more clearly. Other times, it forces me to defend my ideas more rigorously and test their strength. If I’m being completely honest, there are days when I have longer conversations with an AI model than with some people. That’s precisely why, whenever I read or hear a clause like this, I find myself in a slightly uncomfortable position. Instantly, an uninvited meeting begins inside my head. The writer in me insists that creative work must be protected. The AI enthusiast argues that the technology is extraordinary and that its benefits are impossible to ignore. The writer returns to remind everyone that years of work are at stake. The AI enthusiast talks about progress and access to knowledge. And the AI itself, if it had voting rights, would probably announce that it has already processed fourteen million documents while we’ve been having this conversation. At that point, the judge in my head resigns and leaves for an extended vacation.

If this discussion had a moderator, they would probably resign after the first ten minutes too.

– It’s our work, say the authors.

– We need data, say the AI companies.

– We invested in those books too, say the publishers.

– I’m trying to keep up, sighs European legislation.

Meanwhile, the rest of the world is sitting on the sidelines with popcorn, trying to figure out what exactly is happening.

The truth is that, for the first time, I find myself simultaneously in every camp. As a reader, I want access to knowledge. As an author, I want respect for creative work. As an AI user, I see the extraordinary benefits of the technology. As a professional who works with AI, I understand why models need data in order to improve. It’s rather difficult to choose a single team when, in one way or another, you’re playing for all of them.

Of course, authors are not the only ones concerned about this situation. There are also publishers, who view things through a somewhat more pragmatic lens. A publishing house invests in discovering authors, editing, proofreading, design, marketing, and distribution. If authors are the parents of a book, publishers are the relatives who paid for the christening and now want to know who is benefiting from the investment. Which, honestly, is a perfectly reasonable question. Publishers do not operate on stardust, positive energy, and reader appreciation. They have salaries, bills, contracts, and people who need to be paid. From their perspective, if a company develops commercial AI products using millions of books, articles, and publications, it is entirely reasonable to ask who created the raw material and who ultimately benefits from the economic value generated by it.

This is where the great paradox of the entire debate emerges. Many people talk about the issue as if there were two distinct camps: authors and publishers on one side, AI companies on the other. The reality, however, is far more amusing and far less convenient for people who love simple explanations. Many authors already use artificial intelligence. Many publishers already use artificial intelligence. Some publishing houses use AI for marketing, assisted translation, market analysis, or book recommendations. In other words, we are not witnessing a confrontation between the world of books and technology. We are witnessing a negotiation over the rules of the game.

My inner gremlin, however, prefers a much simpler approach.

– Who’s the villain?

– There isn’t one.

– Then who’s the hero?

– There isn’t one either.

– So who wins?

– Depends on who you ask.

– What a poorly organized debate.

And then comes the inevitable question that makes everyone a little uncomfortable. If humans learn by reading books, articles, and studies written by other humans, why should it be entirely different when an AI system learns patterns and relationships from those same texts? Of course, the comparison is not perfect. A human cannot read millions of books in a few months. An AI model can. A human cannot instantly analyze millions of sources. An AI model can. A human cannot process more text in a second than they could read in an entire lifetime. An AI model can. But no creator works in a vacuum. Writers read books. Musicians listen to music. Researchers read studies. Programmers read code. Bloggers read blogs. Civilization itself functions on the principle that we learn from what others have created before us.

Sometimes I’m amused by how quickly we turn artificial intelligence into a digital alien that landed overnight and started stealing ideas from libraries. In reality, much of what it does is exactly what we do: it reads, identifies patterns, makes connections, and generates responses. The difference is that it does all of that in seconds, whereas I require three coffees, two deadlines, and a minor existential crisis. And, if I’m being completely honest, even after all that, I still can’t guarantee the result is any better.

The problem is that the law is trying to determine where inspiration ends and reproduction begins. In Europe, this discussion has already entered the legislative arena. The DSM Directive on copyright in the Digital Single Market permits certain Text and Data Mining activities—that is, the automated analysis of large volumes of text and data. At the same time, however, it gives rights holders the ability to explicitly state that they do not want their works used for that purpose. In other words, authors and publishers can very politely communicate that their work is not an all-you-can-eat buffet for algorithms.

Sometimes I imagine the meeting where someone first proposed adding these clauses.

– We need a clause for AI.

– A clause for what?

– AI.

– You mean robots?

– Yes.

– In books?

– Yes.

– Sounds like the beginning of a science fiction novel.

– I know.

– But this is a legal document.

– I know.

And that is exactly where clauses like this come from:

“No part of this book may be used or reproduced in any manner for the purpose of training artificial intelligence technologies or systems. In accordance with the DSM Directive, [Publisher Name] expressly reserves this work from the Text and Data Mining exception.”

Or, translated into the language we use outside law firms:

“We’re glad you like the book. We really are. But please don’t turn it into algorithm food without talking to us first.”

At first glance, it looks like just another legal paragraph that almost nobody will read. In reality, however, it represents one of the first ways in which the publishing industry is attempting to mark its territory in a world changing faster than legislation can keep up. And that is probably the most fascinating part of the entire story. The success of artificial intelligence does not prove that books have become irrelevant. It proves exactly the opposite. AI models are not trained on millions of pages because those pages are worthless. They are trained on them precisely because they contain the knowledge, experience, culture, language, and creativity accumulated across generations.

My inner gremlin is still trying to simplify the situation.

– So who’s right?

– I don’t know.

– The authors?

– They have good arguments.

– The publishers?

– They have good arguments.

– The AI companies?

– They have good arguments too.

– So what’s the conclusion?

– That this is one of the few debates where almost everyone is a little bit right and almost everyone is also a little bit wrong.

– I hate that answer.

– I know.

Maybe in a few years we’ll have clear rules. Maybe authors and publishers will be compensated automatically. Maybe we’ll find a balance between protecting creativity and developing technology. Or perhaps future generations will look back on this period the same way we look at the early days of the internet and, with the confidence that hindsight always provides, wonder why we argued so much about it in the first place.

Until then, I will continue doing what I usually do: reading books, writing articles, and using artificial intelligence almost every day. Which means that, in this debate, I am simultaneously an author, a reader, and an AI user. Practically speaking, if the war starts, I’ll have to shoot in every direction.

Including at myself. 😏