Artificial Intelligence Will Never Substitute Our Authors. But, Without Regulation, It Will Ruin Publishing As It Exists Today

The single biggest danger to the livelihood of writers and, as a result, to our society is far from being short attention spans. In fact, it is artificial intelligence.

The UK literary sector – contributing more than £11bn – has stood idle while big tech scraped copyrighted material from the internet to refine their models. Recently, an AI startup agreed to a massive legal dispute, but the opportunity has without doubt departed, and big tech is sailing off with the goods.

As a book agent and leader of a major firms in the UK and beyond, I think this is a matter society should be concerned with – not because we fear progress, but because we want to safeguard creativity. By removing the core quality that is intrinsic to being truly human – our ability to think like humans, create stories and conceive alternate realities – we will face a lesser world.

Numerous renowned authors have written about why stories are the lifeblood of humanity and how a writer’s role is to convey insights we might not want to hear. Representing writers such as John le Carré and other literary giants, I know personally where exceptional storytelling comes from.

True authorship is far from being a reproduction of previously published works. It is a blend made up of real-world engagement, faced challenges and understood one’s historical context; it is the outcome of artistry, skill and passion.

The urge to write is not something that can be encouraged – it is an illness that possesses the writer. Truly dedicated writers must write. They might employ editing tools and language models, but nothing would be more abhorrent to an author than a plot being presented to them on their computer that they were then expected to “personalize”.

Some Technology Is Negative

AI that doesn’t replace the creator, or that can collaborate with them honestly, is potentially beneficial. An artist needed for reshoots might allow utilization of existing material to complete a picture. It might reduce expenses, environmental impact and production time. A researcher may wish to streamline their fact-checking by training their personal AI tools. Translation models could broaden the range of global stories, contributing to our literary heritage.

Such uses are worthy of dialogue. But it has to be a debate and remain open to the consumer. Up to now, creative output has simply been stolen without consent, and there are inadequate safeguards on distributors, studios, and industry players.

How Should Take Place?

It begins with some basic principles for everyone to adopt. A protection framework for AI that protects two key elements: permission and attribution.

• No algorithm should be educated with an author’s work without their clear, knowing approval.
• AI firms should be required to make public the content libraries they have used, ensuring clarity so that copyright holders know when their works have been used.
• A writer should also be enabled to withdraw their work easily – not having to discover the choice hidden beneath complex terms of user agreements.

In cases where an artist finds that automated systems is misrepresenting the essence of their output so that it is unrecognizable from the initial version, they should be entitled to withdraw authorization for its application.

Additionally, let’s introduce a tagging mechanism – similar to ingredient disclosures – that prohibit retailers from marketing algorithmically created stories without prominent source identification. Similarly, creator protections must be reinforced, and this can only be done at the legislative tier and even on an worldwide front – a multinational agreement.

Importantly, tech companies should not be enabled to appeal to “fair use” to defend their collecting of protected content. This poses a significant risk to the integrity of artistic rights. It undermines the initial purpose of the “fair use” defense, which was created for educators to quote without fees a small portion from copyrighted material.

These simple guidelines may not seem that important, but they could influence how future generations learn, how historical accounts are preserved, and how we define who we are.