What Audiobooks and Publishing Can Teach Us About the Future of Digital Marketing
Artificial intelligence is steadily making its way into the world of publishing.
It can help prepare a manuscript, check pronunciation, translate content, generate a voice or produce an initial draft.
It can also contribute to the reproduction of existing passages, imitate certain stylistic characteristics and generate books at scale.
This evolution presents many opportunities, but it also raises important questions:
- What happens to creativity when production becomes automated?
- Who is responsible for content generated by a machine?
- Can we truly speak of creation when AI reproduces an existing work?
- What do we lose when the human element disappears from the process?
These questions concern the publishing industry, but they also extend to communications and digital marketing.
AI in the Audiobook Production
Artificial intelligence is already being used at various stages of audiobook production.
It can help to:
- prepare and clean manuscripts;
- identify certain errors;
- check pronunciation;
- translate content;
- create versions in multiple languages;
- generate voices;
- produce dubbed versions;
- adapt content for different markets.
According to a study published in September 2026, 85% of the professionals surveyed already use AI at at least one stage of audiobook production. The applications mentioned include manuscript preparation, pronunciation checks, translation and voice generation.
This figure should nevertheless be interpreted with caution. It does not mean that 85% of audiobooks are produced entirely by artificial intelligence. Rather, it indicates that AI is already present at some point in the production process among the professionals surveyed.
Artificial intelligence can deliver significant time savings for publishers, authors and production studios. It may also make certain projects more accessible, particularly when teams do not have the budget to hire a narrator or book a studio.
But an audiobook is more than a sequence of correctly pronounced words.
A voice contributes to the reading experience. It gives the narrative its rhythm, creates an atmosphere and influences how characters are perceived.
The question, therefore, is not simply: Can a machine read this text?
We must also ask: What does it make the listener feel?
A Voice Is More Than Words
A voice conveys far more than verbal content.
It can express:
- intention;
- emotion;
- personality;
- tension;
- humour;
- a sense of closeness with the listener;
- a particular atmosphere.
Two people can read exactly the same passage and create two entirely different experiences.
One may make it feel warm and intimate. Another may give it a more dramatic, energetic or mysterious quality.
Rhythm, pauses, intonation and vocal variation all contribute to the way a narrative is understood.
In a novel, a voice can make a character seem endearing or unsettling. In a documentary, it can reinforce the credibility of the content. In a children’s book, it can create a particular relationship with the listener.
Technology can reproduce certain vocal characteristics.
But reproducing a voice does not necessarily mean understanding what that voice is meant to convey.
That distinction separates the simple production of sound from interpretation.
When AI Reproduces Existing Works
The debate is not limited to AI-generated voices.
Researchers have shown that, in specific circumstances, some models can reproduce passages that are very close, or even almost identical, to books protected by copyright when prompted in a sufficiently precise way.
This point requires careful qualification.
It does not mean that every response generated by AI is copied from a book. It means that certain models may sometimes memorise and reproduce fragments of texts used during their training.
In some situations, a model may generate:
- passages resembling an existing book;
- wording similar to the original;
- a comparable structure;
- distinctive elements of a work;
- a style strongly associated with a particular author.
These situations are not all legally equivalent. It is important to distinguish between verbatim copying, stylistic imitation, general influence and the reproduction of protected elements.
The essential question, however, remains: At what point does AI-generated output cease to be an original creation and become a reproduction?
Creation, Imitation and Responsibility
AI can be used to generate an idea, suggest a structure or help rework a piece of content.
In such cases, the user remains actively involved in the thinking, selection and rewriting process.
However, AI can also produce text that is very close to an existing work. This may raise questions around plagiarism, copyright and accountability.
It can also imitate the style of a recognisable author or creator. Even where the result is not a word-for-word copy, this kind of imitation can raise ethical and commercial concerns.
The fact that content has been generated by a machine does not automatically make it original.
Nor does it mean that the content is free from copyright restrictions.
The report by the Danish expert group on copyright and artificial intelligence highlights the tensions between models trained on vast quantities of text and the rights of authors, translators and publishers.
AI-generated content should therefore be checked, compared and assessed before it is published or commercialised.
- its originality;
- any similarities it may bear to existing works;
- the data or references used;
- the rights associated with the texts, images or voices;
- how the use of AI should be disclosed.
The Parallel with Digital Marketing
The same issues arise in communication.
Artificial intelligence can generate:
- a blog article;
- a LinkedIn post;
- a slogan;
- a translation;
- a voice-over;
- an image;
- a video;
- an advertising campaign.
However, producing content does not necessarily mean communicating effectively.
A relevant message must answer several questions:
- Who is it addressing?
- In what context will it be received?
- What tone should be used?
- What emotion should it create?
- What action do we want the audience to take?
- What makes the message credible?
- How does it fit within the brand’s identity?
AI can suggest several answers to these questions.
Although it does not always understand the nuances required to choose the right one.
A piece of copy can be grammatically correct, well structured and optimised for search engines while still feeling cold, generic or disconnected from its audience.
The quality of content therefore depends on more than its form.
It also depends on its relevance, context and ability to create a relationship with the audience.
Translating Words or Conveying Intent?
Marketing translation illustrates this difference particularly well.
A direct translation can be grammatically flawless while still being:
- too literal;
- culturally inappropriate;
- overly formal;
- too vague;
- disconnected from the audience’s expectations;
- ineffective in relation to the original objective.
Some expressions cannot be translated word for word.
A slogan may lose its humour. A cultural reference may become incomprehensible. A call to action may sound too strong, or not appropriate in another language.
Transcreation goes further than translation.
It seeks to preserve the intention of the original message while adapting its wording, tone, references and impact for a new audience.
AI can help produce an initial translation.
A person still needs to assess whether that translation feels natural, relevant and consistent with the brand.
The goal is not simply to make the message understood. The goal is to make the audience feel that the message was created with them in mind.
The Genuine Benefits of AI
This is not an argument for presenting artificial intelligence as a threat to every creative profession. Used thoughtfully, it can provide real value.
It can:
- speed up repetitive tasks;
- generate an initial draft;
- suggest different angles or formulations;
- facilitate work across multiple languages;
- identify certain errors;
- help organise ideas;
- test different versions of a message;
- free up time for strategy and creativity.
In a marketing project, AI can be used to prepare a draft, explore possible directions or adapt content into different formats.
It can help a team move more quickly from an initial idea to a first version.
But a first version is not necessarily a final version.
Content still needs to be checked, placed in context, adapted and, in some cases, completely rewritten.
It is also important not to confuse saving time with improving quality.
A tool may accelerate production while increasing the volume of generic or insufficiently reviewed content.
Limitations and Risks
Loss of Authenticity
Content can be produced flawlessly and still feel impersonal.
It may follow grammatical rules, use the right keywords and adopt an effective structure, yet fail to create any genuine emotional response.
If every brand uses the same tools and formulations, their content may eventually begin to sound the same.
A study published in Science Advances showed that generative AI could enhance individual creativity while reducing the collective diversity of creative output.
This finding is particularly relevant to marketing.
AI can help individuals generate ideas more quickly, but when large numbers of creators use the same models, their messages may become increasingly homogeneous.
Contextual Errors
AI can struggle to understand a cultural reference, irony, double meaning or implicit intention.
It may produce a sentence that appears correct but is inappropriate for the situation.
A contextual error can alter how a brand is perceived or make a message difficult to understand.
This risk is particularly significant when content is translated or adapted for multiple markets.
Reproduction of Copyrighted Content
When a model reproduces an existing passage or closely imitates an author’s style, copyright becomes a central concern.
Businesses should avoid publishing AI-generated content automatically, without checking its originality.
They must also pay close attention to the images, voices, texts and references used in their campaigns.
Using AI does not automatically transfer responsibility to the tool.
The organisation publishing the content remains responsible for considering its provenance, compliance and intended use.
The Homogenisation of Messages
AI can produce content quickly, but it can also reinforce the same trends, expressions and structures.
Producing more does not necessarily mean producing better.
Quantity is not a substitute for relevance.
Editorial Responsibility
Even when a tool generates text, an image or a voice, a person must remain responsible for checking, approving and publishing the content.
It is necessary to review:
- the information;
- the references;
- the tone;
- the vocabulary;
- consistency with the brand;
- the originality of the content;
- any potential rights issues. AI can be part of the process. But it should never become an excuse to abandon professional judgement.
Towards Human & AI Collaboration
The future of content will probably be neither entirely human nor entirely automated.
It is more likely to depend on collaboration between technology, expertise and creativity.
AI can take on certain repetitive tasks or quickly produce a range of options.
Professionals can then focus on decisions that require judgement:
- defining objectives;
- understanding the audience;
- choosing the right tone;
- reviewing the results;
- adapting the content;
- protecting the brand’s identity;
- safeguarding creators’ rights;
- making the final decisions.
AI can help produce a first version.
The human being can give it direction.
AI can suggest a translation.
A human can check whether it genuinely conveys the intended meaning.
AI can generate a voice.
A human can determine whether that voice suits the story, the audience and the desired emotion.
Technology then becomes a tool in service of strategy, rather than an automatic replacement for expertise.
Dya Digital’s Perspective
At Dya Digital, we do not see artificial intelligence as an automatic replacement for creativity or human expertise.
We see it as a tool capable of amplifying teams’ abilities, provided that it is guided by a clear strategy.
Used intelligently, it can help organisations to:
- save time;
- explore new ideas;
- adapt content;
- communicate with different audiences;
- improve certain processes;
- develop a more consistent digital presence.
However, technology cannot define a brand’s personality on its own.
It cannot automatically understand every expectation an audience may have.
Nor can it decide, without supervision, what emotion a campaign should create.
It cannot guarantee that content is original or free from copyright concerns.
These decisions require an understanding of the context, the market and the people the message is intended to reach. That is why the use of AI should always be supported by strategic thinking and human review.
Conclusion
Artificial intelligence can generate a voice, but it does not necessarily understand what that voice is meant to make us feel.
It can translate words, but it does not always know how to preserve an intention.
It can produce content, but it does not automatically replace strategy.
It can even reproduce elements of existing works, raising important questions about originality, copyright and responsibility.
In audiobooks and digital marketing alike, the essential question is not simply what AI can do.
We must also ask what it should not do alone.
Because AI can produce content.
But it is the human being who gives it direction, context and meaning.
Sources
ActuaLitté, "Audiobooks: 85% of Professionals Surveyed Already Use AI”, 2026. In the final version, this source should be accompanied by the original study’s methodological details, including the number of respondents, countries covered and definition of AI use.
Le Monde, “How Researchers Managed to Prompt AI Models to Reproduce Books Protected by Copyright”, 2026. This source reports on research concerning the reproduction of passages from copyrighted books under specific conditions.
Ekspertgruppe for ophavsret og kunstig intelligens, report on copyright and artificial intelligence, Danish Ministry of Culture, 2025.
Doshi, A. R. and Hauser, O. P., “Generative AI Enhances Individual Creativity but Reduces the Collective Diversity of Novel Content”, Science Advances, 2024.
French Senate, ChatGPT, et après ? Bilan et perspectives de l’intelligence artificielle, information report, 2024.
