"The crucial question will be: Is this a good listening experience?"

What sets ElevenLabs apart in your view? Why the partnership?

In our view, ElevenLabs is one of the leading technology providers in the field of AI-generated voices. The decisive factor here is not only the sheer quality of the voice, but also its international scalability.

For Bookwire, this is a logical next step: we want to open up additional avenues for publishers to make their content audible, discoverable and monetisable. The aim is not to replace traditional audiobook productions, but to expand the range of offerings – particularly in areas where a full-scale studio production has not previously been economically viable.

How are authors and publishers reacting to synthetic voices? How significant are the reservations?

Reactions vary. There is considerable interest because many publishers immediately recognise that AI audio creates new revenue potential for titles that would not previously have been released as audiobooks at all. At the same time, there are still legitimate reservations – particularly regarding issues of transparency, quality, rights, narrator remuneration and brand impact.

Our experience is that acceptance increases when it is clear that publishers and rights holders retain control, that AI audio is transparently labelled, and that it is not positioned as a substitute for high-quality human productions. Responsible use is key. It must be suited to the title, the genre, the target audience and the publishing strategy.

What has been your experience with the quality of AI-generated audio content? Where does it work well, and where are there problems?

Quality has improved enormously in recent years. With clearly structured texts – particularly non-fiction, self-help books, specialist works, backlist titles or narrative material that is relatively linear – AI audio can already deliver very convincing results today.

We still see limitations with highly literary texts, dialogue, humour, irony, dialects, highly emotional drama or complex character development. A good human narrator interprets a text, consciously employing subtext, rhythm and pauses. AI is increasingly able to simulate this, but cannot always match the standard of a high-quality studio production.

For which titles or segments do you see the greatest added value?

The greatest added value lies initially in the backlist and in titles that have not yet been audiobooked for economic reasons. Of particular interest are non-fiction, self-help books, specialist texts, science, business, spirituality, self-publishing programmes, niche genres and foreign-language rights.

With fiction, it depends more on the individual title. Genre fiction with a clear narrative structure can work. Highly literary titles or works where the narrator’s performance is a central part of the product remain better suited to traditional productions.

Do you envisage hybrid models – such as a human introduction combined with AI for the main text?

Conceivable options include human-spoken forewords, commentaries, interviews or curated framing, combined with AI narration for the main text.

The key will be not to view such formats as a cost-saving measure, but as a new production model with clear quality control.

What new models or formats are conceivable?

If the technical capabilities of platforms are fully exploited, we could see more dynamic audio formats: multilingual versions, personalised voices, chapter-by-chapter updates for specialist books, audio summaries, companion formats, learning versions, etc. could cater even better to user needs.

International rights exploitation is also set to become more exciting: in future, a German publisher will be able to test much more quickly whether a title works as an audiobook in other, even smaller, language markets.

Currently, only around 5 to 10 per cent of books are available as audiobooks. How might this develop?

We expect this proportion to rise significantly in the coming years. ElevenLabs itself states that only around 5 per cent of published books currently exist in audio format – and even fewer in smaller language markets. (ElevenLabs)

By 2030, audio could be a natural consideration for many new releases and large parts of relevant backlists. Not every title will receive a premium studio production, but significantly more titles will be available in audio format at all.

What will the audiobook ecosystem look like in 2030? Will there still be clear distinctions between AI and humans?

If not already the case, by 2030 audio will be an even more integral part of the overall digital publishing ecosystem. The boundaries between e-books, audiobooks, podcasts, summaries and interactive audio may also become more fluid.

Whether users will always be able to – or want to – clearly distinguish whether a voice is human or synthetic remains, to put it mildly, ‘open to question’. But from our perspective, labelling should remain clear. Transparency is key to trust. At the same time, the crucial question will be less ‘Is this AI or a human?’ and more ‘Is this a good listening experience?’

The ability to switch the narrative voice has been shown to prevent listeners from abandoning an audiobook, say the platforms, and that is a good reason to view this technology positively.

Does the audiobook with human narrators still have a future?

Yes, absolutely. Human narrators not only have a future, but remain indispensable for many productions. Human interpretation is a key asset, particularly for literary titles, major authors, well-known voices, children’s audiobooks, elaborately staged productions or emotionally demanding material.

And it is also particularly important that publishers have the choice of which technology they wish to use, or indeed how they wish to produce their content. The possibilities are growing, and the publisher will decide, together with the authors, how their audiobooks are to be recorded or their e-books read aloud – whether using a human voice or an artificially generated one.

Furthermore, we must not forget: consumers have a say in what they like. Questions of quality and whether it suits the product cannot be answered by technology alone, but only through taste, style and craftsmanship.

Nevertheless: AI audio is expanding the market. It does not replace the wonderful art of speech. In the best-case scenario, a larger audiobook ecosystem will emerge: more titles available, more formats, greater international reach – and continued strong, and where appropriate, premium productions featuring human voices where they create the greatest added value.