Blurred lines

The House of Lords recently debated current worries about generative AI raised in the book trade and beyond by the likes of Kate Mosse and Sir Paul McCartney. Baroness Kidron has proposed amendments that would require operators of internet bots that use existing published work to train generative AI models to comply with copyright laws. It seems that the AI rubber is about to hit the publishing road, with radical thinking needed to overcome what is becoming an overwhelming challenge. Here are some lateral ideas designed to initiate conversation in the trade.

With generative AI technology now readily accessible, a surge of AI-generated books is already flooding the market, authored by non-professional writers wielding sophisticated tools to create surprisingly good-quality content. Startups companies Spines, BookBud and Kidzbook and influencers like the Nerdy Novelist cater to this new non-writer market. 

Imagine scrolling through an online bookstore and not knowing if the book you’re eyeing is crafted by a human or synthesised by AI. The line is blurring, and with it arises a compelling question: should we create a distinct category for AI-authored books? Such a move could simultaneously honour the creativity of human authors while embracing AI-generated innovation.

What if the book industry is on the verge of a seismic shift – a new AI-generated format to sit alongside hardbacks, paperbacks, ebooks and audiobooks? It’s a discussion we may need to have sooner than we think.

In recent years, AI has gained traction in areas previously deemed untouchable by machines, including storytelling. Once, a book was a clear expression of human thought and craft, from the author’s spark of inspiration to the final printed word. Now, however, AI can assist or even independently generate entire manuscripts. With programmes capable of producing plot-driven novels, persuasive self-help guides and even poetry, authorship is undergoing a transformation. For non-professionals, AI tools are offering a way to realise literary ambitions previously limited by time, experience or writing ability. These books are landing in online marketplaces and retail spaces, often indistinguishable from traditional titles in format.

The AI rubber is about to hit the publishing road, with radical thinking needed to overcome what is becoming an overwhelming challenge.

The growing volume of AI-generated content brings up some important challenges. For human authors, there is the worry that their work could be overshadowed by AI-created titles. A distinct AI category could alleviate this concern, creating clarity for readers and providing publishers with a means to support the work of human authors. Much like “hardcover” and “paperback” once gave books unique identities within a shared ecosystem, an “AI-authored” label could signal that a book has a distinct origin while allowing it to coexist within mainstream retail.

A real-world parallel can be found in the music industry. As digital compositions became possible, some creators leaned heavily on digital synthesis while others stayed with traditional instruments. Both approaches hold value, but a listener’s experience can differ greatly depending on the production method. Knowing whether a book is AI-generated or human-authored could be similarly significant for readers, who may have preferences regarding their reading experience. For some, the idea of reading a human-crafted story might hold unique appeal; others may be open to AI-authored books or even curious about AI’s creative potential, especially where the price is tempting.

Beyond the marketplace, a separate AI format might shape future publishing norms. Traditionally, book reviews, awards and other forms of recognition have been built around the concept of human authorship. AI-created titles might challenge these conventions if they aren’t differentiated. New standards could evolve, such as awards that specifically celebrate AI-human collaborations or recognise achievements within AI-generated content. Such measures could allow the industry to both welcome AI-driven creativity and protect the distinct craft of human authors.

Furthermore, differentiating AI-authored books can address concerns about market quality and consumer expectations. While some AI-generated books are impressive, others can feel formulaic or lack the nuances typically found in human writing. An “AI” label would give readers transparency, empowering them to make informed decisions about what they are purchasing. Those seeking a traditional storytelling approach may prefer to select books created without significant AI involvement, while other readers might be intrigued by AI’s stylistic quirks and innovations.

Yet, creating a new category for AI-authored books is not without challenges. It could lead to unintended consequences, such as stigmatising AI-created work or creating confusion among consumers unfamiliar with the concept. Publishers would need to establish clear criteria to identify an AI-authored book, such as defining what constitutes a predominantly AI-written versus a human-written piece. These complexities underscore the fact that embracing AI in publishing is as much about responsible integration as it is about innovation.

Is it time for bookstores and publishers to create a unique format for AI-authored books? Acknowledging this category could provide transparency for readers, foster respect for both human and AI-authored content and open doors for a new era of publishing. The conversation is far from over, and the solution may not be simple, but one thing is clear: AI in publishing is here to stay. What do you think – should we create a new category for AI-authored books?