The role of AI in the book publishing industry: a Q&A discussion with George Walkley, publishing AI expert

By   Zoila Marenco 3 min read

George Walkley is an independent consultant, publishing expert and academic researcher. Since 2020, he has worked with over thirty organisations across ten countries. His clients span various sectors, including media and publishing, professional services, education and non-profit.

Q1. Are there any parts of the book publishing industry that you think AI shouldn’t be used or involved in?

AI encompasses a broad range of technologies – it’s so much more than just ChatGPT. I’ve discussed the possibilities with more than a hundred publishers this year and it is clear that AI has use cases across the publishing lifecycle. However, there are two areas in particular that are sensitive for publishers. The first is the distinction between routine and creative work. AI offers great potential for automating tasks and increasing productivity, but there’s real concern around the impact on creative workers, such as authors and illustrators, that has led many publishers to hold back on using AI for content creation. Secondly, generative AI outputs are subject to bias, error and fabrication, so while AI has great potential to help with ideation and drafts, I would not recommend using the output of generative tools without careful human review. For the same reason, I wouldn’t recommend using generative systems for back-office functions such as working with data – however, other technologies such as automation and machine learning could be a gamechanger there.

Q2. What advice would you give to publishing companies on how they communicate this use of AI to their readers?

The key principle is transparency. Many retailers are already asking publishers to label content to distinguish between human creation, AI assistance and AI generation. This highlights that, as well as communicating with readers, publishers need to have conversations with their authors about how they are using AI in their processes.


Q3. Do you think companies that overtly use AI to produce content will be seen negatively by consumers? And what, if any, effect do you think this will have on future consumer behaviour?

I think this rather depends on the type of book. I suspect many people would be disappointed if they found that a novel had been created by or with AI, rather than their favourite author. And any use case where there is an artistic rather than practical element is better suited to human beings – for example, the translation of poetry or literature, where meaning and feeling may be more important than the literal translation. On the other hand, I recently bought an academic monograph which stated on the cover that the translation from the original German edition had been by AI and it didn’t bother me. I would rather have an AI-assisted version than none at all: it’s unlikely that the publisher would ever have offered an English edition of a very specialist monograph if they had to cover the costs of traditional translation. In that context, the combination of AI and print-on-demand probably expands the international market for books significantly.

Q4. What effect do you think AI will have on the imaginative skill of content creators and marketing professionals, both freelance and in publishing houses?

AI is already highly creative and prolific: in a recent study, ChatGPT beat 99% of human participants in a test of creativity. Generative AI is particularly good at generating a large volume of ideas, or creating the first draft of a document. It would be understandable if this worried creatives. But where those people excel is at judging context, refining ideas and using AI as a tool. So I would not position this as human versus AI, but as AI augmenting skilled human judgement. The best creators will be learning how to get the most out of these tools and reinvesting the time saved on manual processes in higher order thinking, in exactly the same way that earlier generations of creatives embraced Photoshop or desktop publishing.

Zoila Marenco
Zoila Marenco
Zoila has five years of experience in client management. She transitioned from working in an organisation offering talent management services to a tech startup specialising in behavioural change in teams. Her experience with clients and communities prompted her move to marketing, taking on the role of a community manager to help Whitefox build, expand and oversee online communities.