Ziya Translation

Half the Publishing Industry Now Works with AI: What the BookNet Report Means for Translation

Published: Updated: 10 min read
Half the Publishing Industry Now Works with AI: What the BookNet Report Means for Translation

The AI debate has moved from the tech pages into the book world's back office — editing, proofreading, and translation desks. A newly published industry study shows that nearly half of the North American book industry now works with AI. In this post, we unpack the BookNet-BISG report's key findings and what they mean for translation work, particularly for certified document translation.

In recent weeks, the artificial intelligence debate has moved beyond the technology pages and into the book's "kitchen" — the editing, proofreading, and translation desks. A newly published industry study shows that the book industry's relationship with AI is no longer in an experimental phase. So what does this picture mean for translation work and for the certified translation of official documents? In this post, we examine the report's key findings, their repercussions in Turkey, and the practical takeaways, especially for readers who need official documents translated.

What Does the BookNet Report Say?

A comprehensive study on AI use across the North American book industry was published as a joint effort by BookNet Canada and the Book Industry Study Group (BISG). According to the report, roughly 48 percent of work in the North American book industry is now carried out with AI support; as industry representatives report, algorithms are involved in many areas — from translation to cover design, from copyediting to page layout — and 48 percent of publishers are actively using this technology. Moreover, contrary to popular belief, this usage is not about making robots write novels; it is preferred for handling the industry's most routine and invisible operational tasks.

The study's methodology is also worth noting: the survey, titled "AI Use Across the North American Book Industry: 2025," collected 559 responses from a wide range of professions and revealed that AI use among North American publishing professionals and librarians is "uneven and mostly exploratory," with cautious optimism coexisting alongside anxiety. While employees at large organizations reported higher usage and greater interest in training, respondents from smaller organizations appeared more hesitant. The literary front of the debate is not empty either: Nobel laureate Olga Tokarczuk's announcement that she used AI during the research phase of her new book was interpreted as a small but deep breach in the walls for those who insist that "creativity belongs to humans alone."

Copyright Concerns: The Industry's Biggest Worry

The strongest concern to emerge in the wake of the report lies on the copyright front. Survey respondents frequently emphasized their concerns about copyright, ethics, employment impacts, environmental costs, and the reliability of AI outputs. These concerns apply to translators as well, because both source texts and translated works are matters of copyright. The translation of a novel or an academic work constitutes an independent work embodying the translator's creative labor; that is why clarifying AI use contractually in literary and artistic translation projects is becoming increasingly critical.

The symbolic example in this debate came from Japan: when Rie Kudan, winner of Japan's most prestigious literary prize, the Akutagawa, admitted that 5 percent of her award-winning novel had been written directly by AI, it caused a major uproar. Another important finding of the report concerns the regulatory gap: while AI adoption is advancing rapidly, the industry's lack of consistent standards, governance mechanisms, and shared best practices stands out. In other words, the technology is running ahead of the rules — which means legal uncertainty for both publishers and translators.

Is Translation and Editing Labor at Risk in Turkey?

The real question being debated in Turkey is this: will Turkish publishers, already strained by the economic crisis, also entrust editing, proofreading, and translation labor to algorithms in the name of "cutting costs"? In this scenario, quantity may increase and shelves may fill with faster, cheaper books; but where the quality of language and the standard of publishing will end up remains uncertain. Indeed, critics point to the same issue: AI can find grammatical errors or generate translation drafts; but grasping a text's spirit, cultural codes, and emotional depth always requires the human touch.

Although book translation and official document translation are different fields, the message is the same: AI speeds up production, but what guarantees quality is still skilled human labor. This balance is especially evident in academic translation; in texts such as theses, articles, and diploma supplements, terminological consistency and knowledge of institutional formats are added values that only a subject-matter expert translator can provide. An academic who settles for machine output may face a serious credibility problem in the eyes of a journal reviewer or an equivalency authority.

What Can AI Do in Translation — and What Can't It Do?

The framework drawn by industry representatives is quite clear: AI cannot fully replace an author or a senior editor, but it serves as a serious assistant in the book's kitchen. In translation specifically, this means speed gains in tasks such as producing first drafts, suggesting terminology, checking consistency, and pre-screening high-volume texts. But the picture changes for legally binding texts: a single wrong term in a contract can alter the parties' obligations, which is why human expertise remains indispensable in fields such as legal translation.

The same sensitivity applies to the medical field: in documents such as discharge summaries, laboratory results, or drug package inserts, a nuance missed by machine translation can directly affect patient safety. That is why double-checking (translator + editor) is standard practice in medical translation projects. The table below summarizes where three different translation approaches stand:

FeatureMachine / AI TranslationSworn TranslationNotarized Translation
Official validityNoneSufficient for many institutionsRequired by institutions demanding notarization
LiabilityUnclearUnder the sworn translator's signature and stampTranslator's signature + notary certification
Typical usePreliminary reading, drafts, internal correspondenceDiplomas, transcripts, institutional applicationsPowers of attorney, court documents, corporate papers
CostLowTranslation feeTranslation fee + notary fees and charges

The Line Is Clear for Official Documents: Humans Have the Final Say

Sworn translation is of critical importance in situations where official documents must be translated accurately and reliably. AI can lighten the workload by preparing initial drafts of documents; however, final approval and verification must be performed by a sworn translator, because even the smallest error in such documents can have serious consequences. The principle critics articulate for publishing applies here word for word: even though the adoption rate is a big step for the industry, humans must always have the final say. At sworn translation offices such as Ziya Tercüme, technology is positioned alongside human oversight, not in place of it: for documents such as diplomas, court rulings, powers of attorney, or medical reports, a sworn translator handles the translation, signs it, and bears the legal responsibility.

At this point, a frequently confused distinction also needs clarifying: sworn translator certification and notarization are not the same thing. A sworn translator attests to the accuracy of the translation with their own signature and stamp; a notarized translation involves this translation additionally being certified by a notary and is subject to an extra notary fee. If the document needs to be valid abroad, a third layer comes into play: the apostille. A translation office cannot issue an apostille; in Turkey, this annotation is issued, depending on the type of document, by authorized administrative authorities such as governorships/district governorships (valilik/kaymakamlık) or by judicial authorities. In apostille-certified translation processes, the office's role is to prepare the translation and the certification chain in line with the destination country's requirements.

Sample Scenario: Diploma Translation for Study Abroad, Step by Step

To make the report's abstract findings concrete, consider a typical scenario: a student applying for a master's program in Germany needs to prepare their bachelor's diploma and transcript. In this process, AI's place is at most the student's own preliminary reading; the path to follow for the copies that will go into the application file is well established:

  1. Confirm the target university's document requirements (language, level of certification, apostille requirement) in writing.
  2. Prepare the originals of the diploma and transcript, or institution-certified copies.
  3. Have the documents translated by a sworn translator; the translator signs and stamps the translation.
  4. If the institution requires notarization, have the translation certified at a notary.
  5. If an apostille is required, obtain the apostille annotation from the competent authority (for administrative documents, the governorship/district governorship).
  6. Scan all copies and keep digital versions; many universities ask for electronic upload first.

The most common mistake in this chain is breaking the sequence: for instance, which document the apostille applies to (the original, the translation, or the notarization) varies depending on the target institution. Another trap is trying to present a machine-translated text as "sworn"; translations without a signature and stamp will not be processed by official authorities and can cause delays in the application timeline that are hard to recover from.

Factors That Affect the Cost

In cost planning, two separate items must be distinguished: the translation fee and the notary fee. Under the legislation in force as of this article's publication date, pursuant to the 2026 Notary Fee Tariff, a translation fee of TRY 667.67 is charged for each page of documents translated from one language to another by or through a notary; for pages containing 10 lines or fewer, half of this fee applies. The tariff took effect on January 1, 2026 and was published in the Official Gazette (Resmî Gazete) No. 33125, dated December 30, 2025. The translation fee — the translation office's labor — is separate from this amount and varies depending on the language pair, document density, and turnaround time.

The main variables affecting total cost are: how common the source and target languages are (there can be a marked difference between English and a rare language), page/character volume, terminology density (legal and medical texts require more expertise), the certification layer (sworn only, notarized, or apostilled), and urgency. An exact figure requires seeing the document; that is why requesting a written quote based on the document is always more reliable than "ballpark" prices given over the phone.

Practical Notes for Document Holders

In light of the report and current debates, we have compiled the key points for readers who will use translation services in the period ahead. These items serve as a checklist for both individual document holders and companies with regular translation needs:

  • In publishing and academic texts, AI-assisted drafts can speed things up; however, every text to be published or submitted to an official authority must pass through human editor/translator review.
  • In official procedures (equivalency, marriage, company formation, court proceedings), institutions do not accept machine translation; a signed and stamped sworn translation is required.
  • When planning costs, remember that the translation fee in the 2026 notary fee tariff is TRY 667.67 per page and that the translation office's fee is calculated separately.
  • For copyright-sensitive texts (books, articles, catalogs), request an explicit clause on AI use in your translation contract.
  • Plan the certification chain (sworn signature → notarization → apostille) from the outset according to the target institution's requirements; a sequencing error can send the whole process back to square one.

The Industry's Next Step: The 2026 Survey and the Search for Standards

This debate is far from closed; on the contrary, it is becoming institutionalized. BISG and BookNet Canada announced plans to field the survey again in 2026 to track, on a data-driven basis, how adoption rates and industry opinions change over time. Indeed, the second "AI in the Book Industry" survey went live this summer; open to all branches of the industry — including publishing, distribution, printing, libraries, and freelance professionals — the survey collected responses until July 17. When the results are published, we will see clearly how adoption has shifted within a year.

In short, contrary to fears, the report does not paint a picture in which "humans are cut out of the loop"; it points to an order in which AI takes on the drudge work while the value of skilled language labor increases. In official document translation, this balance is already a legal requirement: the final word — and the signature — always belongs to the sworn translator. If you are unsure which certification layer your document needs, confirming the current requirements of the institution where you will submit it and planning the process accordingly will save you both time and money.

Sources

Professional Translation Services

Sworn, notarised and apostille translations in 50+ languages. Fast, reliable, enterprise-grade solutions.