More than half of published novelists in the UK believe artificial intelligence could eventually replace their work entirely, according to a new report from the University of Cambridge.
The study, conducted for the university’s Minderoo Centre for Technology and Democracy, suggests widespread unease about the speed and scale of AI’s advance into the literary world.
The research surveyed 258 published novelists and 74 industry figures including editors and agents. Just over half (51%) of novelists said that AI is likely to end up entirely replacing their work. Many participants reported that their work had already been used without their permission to train large language models, and more than a third (39%) said their income had fallen as a result of generative AI. A large majority also expected their earnings to decline further.
“There is widespread concern from novelists that generative AI trained on vast amounts of fiction will undermine the value of writing and compete with human novelists,” said Dr Clementine Collett, author of the report.
“Many novelists felt uncertain there will be an appetite for complex, long-form writing in years to come. Novels contribute more than we can imagine to our society, culture, and to the lives of individuals,” she added.
Tracy Chevalier, author of Girl With a Pearl Earring said: “I worry that an industry driven mainly by profit will be tempted to use AI more and more to generate books.
“If it is cheaper to produce novels using AI … publishers will almost inevitably choose to publish them. And if they are priced cheaper than human-made books, readers are likely to buy them, the way we buy machine-made jumpers rather than the more expensive hand-knitted ones.”
Romance authors were identified as the most exposed to displacement by AI tools now capable of producing long-form fiction, followed closely by thriller and crime novelists. Many respondents described the market as increasingly crowded with AI-generated books, with some saying they had discovered titles listed online under their name that they had not written. Others reported reviews that appeared to be generated by AI, featuring muddled characters or incorrect details, which they feared could harm sales.
The report comes amid increased scrutiny of AI-generated content on online retailers such as Amazon’s marketplace, which experts have warned are a “wild west” due to the lack of regulation of AI-generated books.
Despite this, the report found that attitudes towards AI were not uniformly hostile. Around a third of novelists said they already used AI, mainly to carry out tasks such as sourcing information. However, almost all expressed strong opposition to AI writing novels or even short passages of text, and many rejected the idea of using AI for editing.
A recurring theme in the research was the sense that copyright protections had failed to keep pace with technological change. Authors said they wanted informed consent and payment for any use of their work in AI training, along with greater transparency from technology companies and support from the government. The report notes frustration with the government’s proposed “rights reservation” system that would allow AI firms to mine text unless authors opt out.
Authors fear AI may weaken the deep human connection between writers and readers at a time when reading is already at historically low levels, particularly among children: only a third of UK children say they enjoy reading in their free time, the lowest level in two decades.
The report also comes at a moment of particular tension between authors and AI companies. In September, Anthropic agreed to pay $1.5bn (£1.1bn) to authors who accused the company of using pirated copies of their works to train its chatbot.
