Audible is being sued in the USA by seven major publishers over potential copyright infringement resulting from their proposed new feature ‘Audible Captions’, software utilising AI to transcribe the spoken word.


The software was announced at the end of July 2019, due for release on 10th September. The feature uses machine learning to turn speech to text in real time, which would enable users who struggle to listen to be able to read segments on-screen along with the audiobook.

It is marketed largely as an educational tool but can benefit those who are hard of hearing or people looking to learn a language.

However, the initiative’s launch has been halted following a copyright suit brought in the Southern District of New York by seven of the largest publishers in America: Chronicle Books; Hatchet; HarperCollins; Macmillan; Penguin Random House; Scholastic and Simon & Schuster.


The complaint asserts that Audible’s features show a disregard for various provisions of the Copyright Act 1976 such authors’ and publishers’ exclusive rights to control and financially benefit from the distribution of texts for which they have the copyright.

The intention of this, as argued in the complaint, is “to cut Publishers out from a business model that already exists, by unlawfully creating derivative works of, reproducing, distributing, and publicly displaying unauthorized copies of the Works” [3].

The distribution of works without the publishers consent may also be in contravention of Article 6(1) of the WIPO Copyright Treaty of 1996) in that authors “shall enjoy the exclusive right of authorizing the making available to the public of the original and copies of their works through sale or other transfer of ownership”.

All seven publishers are members of the Association of American Publishers (AAP) and, as joint plaintiffs, filed for an immediate injunction to prevent the books for which they own the rights from being included in Audible Captions. The claim is backed by the AAP and The Authors Guild.


The publishers note that the necessary licenses for use of books as textual and audio versions differ, and that Audible intends to create, copy and distribute textual versions “despite not holding the rights to do so” [3], which would offend their copyright

Similarly the creation has, the publishers argue, already occurred on a some scale as without testing the software on published texts for which they do not hold the copyright, Audible would not have been able to say that up to 6% of AI transcription could have errors. Therefore, there is not only a potential breach, but a breach has already occurred.

The electronic rights of e-books belong exclusively to the publishers and so error riddled AI transcription is not only a quality concern but can be seen as way of circumventing and undermining these rights. This could have grave ramifications for the ebook production industry on a broader scale.

They also state that Audible “does not plan to compensate Publishers or their authors for this feature, nor will it allow them to decide what titles will be made available as Distributed Text” [3].

In a statement on the AAP website from 23 August 2019, the publishers argued that Amazon’s move was an effort to seek commercial advantage from literary works that it did not create and does not own, Audible is wilfully pushing a product that is unauthorized, interferes and competes with established markets”.


Audible has publicly refuted AAP’s arguments in an online statement, claiming that since the transcriptions are not whole textual reproductions but small sections of text appearing and disappearing in time with the audio, and that the revelation of text is very controlled, this means that it does not constitute the creation of a copy of a book and therefore there is no need to obtain any further licences .

In their statement responding to the AAP Audible wrote that it blatantly “disagree[s] with the claims that this violates any rights and look forward to working with publishers and members of the professional creative community to help them better understand the educational and accessibility benefits of this innovation“.

However, the definition of a reproduction under the Copyright Act is the copying of either an ‘entire work, or to a substantial part of it’ – therefore even if it is not all on the screen at once, the creation and distribution of a substantial part of the work can occur through audible captions in a very short time period and to many people at once.

Additionally, the argument that Audible’s use of the texts constitutes ‘fair use’ is debatable as it engages with several of the factors in paragraph 107 of the Copyright Act, arguably most significantly factor 4 – the effect of the use upon the potential market for or value of the copy – Audible’s free versions would likely devalue those of the publishers as a reader would be more inclined to access a free copy than one that costs.

Audible have until 13 September 2019 to file a response.


The most similar feature to Audible Captions is Immersion Reading, an Amazon feature which allows a user to listen to an audiobook on Audible whilst reading the e-book on Kindle.

However, as stated at point 38 of the AAP statement it is in practice very different as it requires the user to have purchased both formats separately. Publishers agreed to this feature when it was released as they benefit financially from the use of both versions, each version being obtained from a licensed distributor.


Audible filed a stipulation on 28 August, conceding that until any proceedings are finalised it will not make Audible Captions available for books from those publishers.

Any potential ruling on this issue would likely impact the feasibility of the Audible Captions feature on a broader scale and future AI transcription of published works. The injunction hearing is currently scheduled for the end of September.