The publishing sector has been open to valid criticism for its failure to quickly adopt the kind of culture and business change technology that has delivered massive benefits for industries such as music and ecommerce. In particular there has been significant fear, uncertainty and doubt around embracing AI in publishing and the potential solutions, mainly around the issue of creative integrity.
AI is not one thing
The notion of technology adoption is a complex one and the area of AI is really a bundle of different technologies, some innovative and fresh to the market, others rebranded to take advantage of the boom in interest in the AI brand, that need a good deal of unpacking, rather than being a single, unified whole. In some ways, the latter point has been part of the source of the confusion.
It is worth being specific about what that bundle actually contains. Machine translation, generative voice synthesis, recommendation algorithms, and large language models all travel under the same “AI” label, but they are doing very different things inside a publishing business. An AI in publishing tool that helps a rights team identify foreign language markets for a backlist title is a different proposition entirely from one that drafts cover copy or generates synthetic narration. Treating them as a single category has made the conversation harder than it needs to be, and has sometimes led to blanket scepticism where more selective engagement would have been more productive.
What mainstream adoption looks like
But however the publishing sector, and the wider general public, has grappled with the ideas of AI, we are starting to see some good examples of adoption in safe and sensible ways.
Aside from those who are pioneering the boundaries of what digital narratives look like, the more mainstream operators are starting to think about more effective technology partnerships. If we look at less contentious partnerships we can identify ones such as when Penguin Random House has begun working with social media platform TikTok to expand out the #BookTok functions to further connect with readership in the digital space where they spend their time.
Further to this, in recent years several publishers have developed and worked with strategic partnerships to test and integrate AI tools and workflows. We see this with Pan Macmillan’s work with CHAPTR, Harlequin France’s trial of AI translation, and with some of the biggest publishers hiring AI engineers.
The translation example is particularly worth dwelling on. The economics of literary translation have long meant that only a fraction of the world’s published titles ever reach readers in other languages. AI-assisted translation does not replace the literary translator, but it is beginning to make the process faster and more cost-effective in ways that could meaningfully widen the range of titles that cross language barriers.
The audiobook revolution
And perhaps one of the best areas of growth is around the audiobook sector, where generative voice technology has led to some interesting products and services becoming available for use and development. For example, Spotify has recently launched an audiobook creation product, working in partnership with ElevenLabs, alongside publisher partnerships with the platform including HarperCollins, Storytel, and Bertelsmann.
The significance of this development should not be understated. Audiobook production has traditionally been expensive and time-consuming, which has meant that a large portion of published titles, particularly from independent and mid-size publishers, never made it into audio at all. Generative voice technology is beginning to change that equation. The quality gap between AI narration and human performance remains a genuine consideration, and listener preferences vary considerably, but the accessibility gains are substantial: books that would never have reached blind or visually impaired readers, or readers with dyslexia, are now becoming available in audio formats that simply would not have existed otherwise.
What’s particularly heartening is that much of the work in this field seems to be oriented around broadening the accessibility of books in all of their forms to new and existing readers. Services such Bookshare and Dudley Editions seem well placed to build and broaden the appetite for digitally enabled storytelling services.
Getting AI in publishing right
As with all digital change programmes, the challenges remain continuous and ongoing. Questions of authorship and attribution are still being worked through legally and ethically, particularly as AI tools become more embedded in the writing and editing process. The compensation question for voice artists and translators is a live one: the efficiency gains of generative technology are real, but so is the risk that they come at the cost of livelihoods that have historically been precarious. There is also a subtler concern worth naming: that recommendation algorithms, however sophisticated, tend to surface what is already popular, which can quietly narrow the range of work that reaches readers. None of these are reasons to halt adoption, but they are reasons to keep human judgement at the centre of the decisions being made.
That well-informed standpoint has institutional backing. The Publishers Association in the UK has issued guidance on AI in publishing’s use that attempts to distinguish between tools that support human creativity and those that risk displacing it. In the US, the Association of American Publishers has been similarly active in engaging with policymakers on copyright and licensing questions that AI raises for the industry. These are not perfect frameworks, and the pace of technological change means they will need continuous revision. But their existence signals that the sector is at least attempting to govern its own adoption thoughtfully, rather than leaving it entirely to platform and technology companies to set the terms.
What’s more certain is that people seem to be approaching things from a well-informed standpoint, keen to make best use of the technology while focusing their efforts on improving things for their readerships and workforces alike.