Wie KI jedem Hörbuchhörer weltweit personalisierte Erzählungen bietet

Imagine you’ve been waiting anxiously for the sequel to your favourite book. Finally, after months, it’s here. You download the audiobook, and press play. In your head, you’d imagined the narrator with a warm, gravelly tone — maybe a hint of a posh British accent. But instead, the voice is flat, or too high-pitched, or just wrong.

With a physical book, the possibilities are unlimited: your mind shapes the characters’ voices. With traditional audiobooks, that choice is made for you. For decades, audiobooks have been limited to a single voice, dictated by publishers’ decisions, budget constraints, and the availability of narrators. But AI is changing that. Today, you don’t just have to passively listen, you can actively engage with the book in a way of your choosing.

This shift matters because the audiobook market is growing at an unprecedented rate. By 2029, the number of audiobook listeners is expected to reach 1.8 billion. Global revenue is also climbing, projected to hit $9.84 billion by the end of 2025 and surpass $13 billion by 2029.

Despite these gains, the current market figures primarily reflect traditional audiobook production, which remains limited by cost and narrator availability. We believe voice AI will accelerate this growth by reducing production barriers, making audiobooks more accessible across languages and regions, and enabling more creators to bring their stories to life. 

As demand for immersive, convenient storytelling continues to rise, AI-driven narration will help unlock the full potential of the audiobook market.

Beyond one voice: Personalization through AI

Traditional audiobooks have long been limited to a single narrator chosen by publishers — often based on budget, author input, and availability rather than listener preference. An English-language audiobook might feature an American voice, but an Australian or Indian listener may prefer a narrator with a more familiar accent.

Unlike reading a physical book — where readers naturally imagine character voices — traditional audiobooks don’t offer that flexibility. As New York Times best-selling author Joseph Nassise points out:

Everybody's different. What I may like or imagine as an author might be different than what listeners may actually prefer. It’s powerful to give my listeners the flexibility to choose how they want the story told.

AI-driven narration changes this by allowing users to customize their listening experience, making audiobooks more immersive and accessible.

Beyond personal preference, AI narration removes barriers faced by authors, particularly indie writers and those from smaller markets. The cost and production time of human narration have traditionally been prohibitive, and even authors who want to narrate their own work don’t always get selected.

As USA Today best-selling author Leeanna Morgan notes:

For the fist time, readers are in control of how they want to hear their audiobooks. The ElevenReader app allows listeners to customise their audiobook experience and bring an author's stories to life in the most meaningful way for them.

The ElevenReader app already enables hundreds of thousands of users in over 32 languages to select their preferred AI voice — with voice actors also compensated, paying out over $1 million in earnings last year alone.

Language selection has also been a major limitation in traditional audiobook production, with most recordings limited to globally dominant languages. Even when translations exist, localized accents and dialects are rarely considered.

ElevenLabs' voice technology changes this by generating highly localized voices with contextual awareness — ensuring narration sounds natural and emotionally resonant. Our data from nearly 900,000 hours of audiobook listening shows that users strongly prefer localized voices. 

Across the eight largest global languages in the ElevenLabs reader app, over 34 unique AI voices are represented, with the most popular voice in any given language accounting for just 15% of total listening.

This diversity highlights the demand for narration that feels native to each listener. Even within the same language, preferences vary widely.

In English, for example, Brian (American accent) is the top choice in the U.S., while George (British accent) is preferred in the UK and Australia. A similar trend appears in Portuguese — Jessica is the most popular voice in Brazil, while European Portuguese listeners prefer David.

The data reveals that only one AI voice was shared among the top five in both Portugal and Brazil, despite their shared language roots. This demonstrates the importance of regional variety. While this preference for dialect seems intuitive, tone of voice is even more subjective.

A thriller set in Glasgow may sound best with a Glaswegian accent, while a Parisian love story feels more natural with a French one.

Our analysis of nearly 900,000 hours of audiobook listening confirms that no single voice fits all listeners. Even within one language, AI voices display a "long tail" of user preference — where the top narrator in a sample represents just 10% of listening, requiring more than a dozen voices to reach 50% of usage.

This reinforces a reality that traditional audiobook production simply cannot match: listeners around the globe want a choice, and AI makes that possible.