In 2017, Vladimir Putin warned that artificial intelligence would determine global power: “Whoever becomes the leader in this sphere will become the ruler of the world.” Nearly a decade later, the publishing industry views the same technology with suspicion. This reaction raises a deeper question about the very purpose for which the technology was created.
On April 10, The New York Times reported that Hachette Book Group cancelled the U.S. release of the novel Shy Girl over concerns about possible AI involvement. Readers felt misled, authors felt exposed, and publishers appeared unprepared. The episode has been framed as a crisis of technology. It is not. It is a crisis of definition. The debate has settled on the wrong question: Was AI used? The more fundamental question is: What constitutes authorship?
Much of the current panic rests on detection tools whose unreliability is already well documented. Antonio Bricio, a debut writer, submitted a chapter of his novel to a detection system that declared it 100 per cent AI-generated. After minor edits, the same system judged it entirely human.
Andrea Bartz encountered a similar problem when her own writing was flagged as largely AI-produced. These are not anomalies. They reveal a deeper problem: the industry is attempting to police authorship with instruments that cannot measure it. Honest writers risk false accusation, while the real question of whether the ideas themselves are original remains unexamined.
The physical act of writing has never defined authorship. The origin and control of ideas has always defined it.
Every book that reaches a reader is already the product of collaboration. Developmental editors restructure narratives. Line editors refine arguments. Copy editors polish language. In some cases, ghostwriters produce entire manuscripts. Yet the work is attributed to a single author because that author is the intellectual source and final authority.
Now replace the editor with AI, and the same collaborative process becomes suspect. This is an inconsistency that publishing cannot sustain. If external assistance invalidates authorship, then much of modern publishing fails its own test. If authorship resides in the origin of thought, then the nature of the tool, human or machine, is secondary.
The confusion has reached an almost absurd extreme. Writers are now placing logos on their books declaring their work “human-authored” — a label that would have been incomprehensible to any previous generation. No author ever felt compelled to certify that a typewritten manuscript was humanly conceived, even after it had passed through layers of editorial intervention. It is a sign of how uncertain the industry’s thinking has become. The real distinction is not between AI use and non-use. It is between assistance and replacement.
When AI functions as an extension of the writer, gathering data, translating language, and organising material, it operates no differently from a library, a calculator, or an editor. The mind directs; the tool executes. When AI replaces the writer’s thinking, constructing arguments, shaping narratives, and producing prose with minimal human intervention, the situation changes.
Authorship does not reside in the instrument. It resides in the mind
What is lost is not craftsmanship but intellectual ownership. Even then, the standard must remain clear: whether the author is the originator of the ideas or merely presenting generated output. Plagiarism was a violation before AI arrived. It remains one regardless of the instrument used.
In my own work, I use AI to gather data, locate official reports, and convert information into usable formats. When writing on flooding in Pakistan or water scarcity along the Indus River, it helps assemble datasets. When covering education, it extracts tables from official reports. It translates drafts into Urdu, which I then rewrite and refine. None of this produces the argument itself. The ideas, the structure, and the interpretation remain entirely my own. The tool accelerates the process. It does not replace the thinking.
A useful parallel comes from education. Two models of examination exist. The memory-based model tests a student’s ability to reproduce stored information. The open-book model allows students to consult texts and find relevant material. Cheating concerns persist in closed examinations, not open ones.
When resources are freely available, copying becomes pointless — only original thinking earns the marks. The book is a tool; the thinking is the test. AI in writing works identically. A columnist who uses AI to gather data and build an original argument has passed the open-book test. A novelist who feeds a premise into a chatbot and publishes the output has simply copied from the book.
Some argue that generative AI represents a genuinely new category — the first tool that produces content rather than merely processing or reproducing it. This is true and deserves acknowledgement. But producing content that resembles original thought is not the same as originating thought. AI recombines existing human knowledge at extraordinary speed. It does not create from nothing. The philosophical principle remains intact: what distinguishes authorship is not the fluency of the output but the origin of the ideas behind it.
This anxiety follows a recurring pattern in the history of tools. Even in the Ottoman world, manuscripts remained preferred over printed books for generations, reflecting a deeper tendency: new technologies are resisted not because they fail, but because they unsettle established ways of thinking. In Pakistan, sections of the clergy once condemned loudspeakers and television, only to later chase airtime on every channel.
Every tool begins as a threat, then becomes indispensable. If we accept the logic that AI taints whatever it touches, we must follow it to its conclusion — that every tool which makes writing easier, research faster, or thinking sharper is a corruption of authentic human effort. That conclusion does not protect creativity. It imprisons it.
Every transformative tool has also democratised the field it entered. The printing press expanded authorship beyond monasteries. The computer made writing universally accessible. AI is extending this pattern. A writer in Karachi, Lagos, or Guadalajara can now access editorial and research assistance once restricted to those with institutional backing. When tools become widely available, ideas become the only remaining differentiator. Resistance to AI, in this context, is not only about authenticity. It is also about reflecting concerns about gatekeeping as much as about standards.
On the question of transparency: voluntary acknowledgement of significant AI assistance, following the same spirit as editorial acknowledgements that already appear in most books, is a reasonable and honest practice. Many authors already thank editors who restructured entire chapters.
Acknowledging AI assistance in a preface deserves neither more nor less than that. But making labelling mandatory or cancelling a book’s release over unverified detection results is a different matter entirely. Mandatory labels, moreover, can mislead a reader who sees “AI-supported” and may assume “fully AI-generated”. This is not an argument against transparency. It is an argument against substituting labels for judgement.
The task before publishing is not to police tools with unreliable detectors or reduce authorship to disclosure labels. It is to clarify what authorship has always meant. A fully disclosed machine-generated text does not become authored simply because it is labelled. A genuinely conceived work does not lose its authorship because a tool assisted its execution. Authorship does not reside in the instrument. It resides in the mind.
Once that is understood, the panic surrounding AI will follow the same path as every technological disruption before it — moving from suspicion to acceptance, from heresy to orthodoxy. From the typewriter in Istanbul to the loudspeaker in Pakistan, the pattern has never changed.
Nearly a decade after Vladimir Putin framed artificial intelligence as a source of global power, the publishing industry is still struggling to understand its place in something far more basic: authorship. The question is no longer whether AI will shape the future, but whether we understand the principles that should guide its use.