Half of surveyed UK novelists fear AI could replace their work entirely

For many British novelists, the threat from artificial intelligence reaches beyond the manuscript on their desk. It also touches the freelance assignments that pay their bills, their ownership of published work, and their connection with readers.

A University of Cambridge report found that 51% of participating published novelists believed AI would likely replace their fiction-writing work entirely. Another finding was more immediate: 39% said generative AI had already hurt their income. Some 85% expected their future earnings to decline because of it.

Published in, The Impact of Generative AI on the Novel examined experiences across British fiction. Clementine Collett, a BRAID UK Research Fellow at Cambridge’s Minderoo Centre for Technology and Democracy, led the research. The report was published in association with the Institute for the Future of Work.

The survey gathered 332 responses between February and May 2025. Participants included 258 published novelists, 42 fiction-publishing professionals, and 32 literary agents. Among the novelists, 233 were traditionally published and 25 were self-published.

Collett also conducted focus groups and interviews, alongside a Cambridge forum involving writers, publishers, and other stakeholders. These discussions explored experiences behind the percentages, including concerns about contracts, creative practice, and competition.

The findings describe participants’ reported experiences and expectations. They do not establish how many writing jobs AI will eliminate or independently measure the income losses attributed to it. The survey sample was also predominantly white and female, with 87% and 78% of respondents identifying that way, respectively.

“Many novelists felt uncertain there will be an appetite for complex, long-form writing in years to come,” Collett said.

Some writers described competition from AI-generated books in an increasingly crowded market. Others reported finding books listed under their names that they had never written. Participants also raised concerns about apparently AI-generated reviews containing confused character names and damaging ratings.

Those accounts point to several different pressures. A novelist may face competition for readers while also losing the separate work that makes writing financially possible.

“Most authors do not earn enough from novels alone and rely on income streams such as freelance copywriting or translation which are rapidly drying up due to generative AI,” Collett said.

The report also recorded fears of a divided market. Some participants envisioned human-written novels becoming expensive luxury products while AI-generated fiction became cheap or free. That was a possible future described by respondents, rather than a market outcome established by the research.

Across all surveyed literary professionals, romance was considered especially vulnerable. Some 66% described romance authors as “extremely threatened,” followed by thriller writers at 61% and crime writers at 60%. Literary fiction received that rating from 32%.

These assessments reflect perceived exposure to displacement. The report did not test AI novels against human books or compare actual job losses across genres.

Participants worried that wider AI use would make fiction less original in its content, language, and style. They also feared output could repeat stereotypes drawn from existing material. Some suggested writers might respond with more experimental work, pushing their art in directions they believed machines would struggle to follow.

For others, the concern centered on the process itself. Writing and revising were described as ways of developing imagination, judgment, and an understanding of human complexity. Removing difficulty from that process could, participants feared, weaken what eventually reached readers.

The responses did not amount to blanket opposition to artificial intelligence. Some 80% of all respondents agreed that AI offered benefits to parts of society. A third of novelists reported using it in their work, mainly for tasks they considered non-creative.

Around 20% used AI to source general facts or information. About 8% used it to edit text they had written without AI. Editing remained contentious, however, with 43% feeling “extremely negative” about using AI for it.

Resistance grew much stronger when the technology supplied prose. Some 97% of novelists felt “extremely negative” about AI writing entire novels, and 87% felt that way about short sections.

Transparency was another central concern. Writers feared undisclosed AI use could erode trust with readers, while false accusations could damage an author’s reputation. Participants wanted clearer information about how publishers and writers used these systems.

Some 59% of novelists reported knowing that their work had been used to train generative AI. Among that group, 99% said they had not given permission, and all said they had received no payment.

The survey found strong resistance to a proposed copyright exception allowing text mining unless rights-holders opted out. Some 83% of all respondents considered that approach negative for publishing. Among novelists, 93% said they would probably or definitely opt out.

Instead, 86% of all respondents preferred an opt-in principle requiring permission and remuneration. Almost half of novelists, 48%, favored licensing negotiated collectively through a writers’ union or society.

The policy context has since changed. In its March 2026 copyright and AI report, the UK government said a broad exception with opt-out was no longer its preferred approach. It proposed gathering more evidence and considering other policy options.

Collett’s recommendations include greater transparency about training data, fair payment, and support for licensing. The report also calls for clearer guidance on AI rights in publishing contracts, particularly for writers without agents.

Beyond copyright, participants wanted support for independent publishers and writing initiatives involving underrepresented voices. They also advocated creative writing without generative AI in schools, while recognizing accessibility needs, alongside teaching that helps children assess the technology critically.

The report presents those measures as ways to protect both livelihoods and literary variety. Its central question is how writers can retain control of their work while continuing to develop it.

“The novel is a precious and vital form of creativity that is worth fighting for,” Collett said.

These resources examine creative diversity, employment exposure, copyright policy, and the reading habits that shape fiction’s future.

Generative AI enhances individual creativity but reduces the collective diversity of novel content: An experiment examines how AI-generated ideas affect short-story quality and similarities across writers’ work. (Science Advances, 2024)

Generative AI and Jobs: A Refined Global Index of Occupational Exposure: This task-based assessment distinguishes potential exposure to generative AI from actual employment losses. (International Labour Organization, 2025)

Report on Copyright and Artificial Intelligence: The government reviews consultation responses, licensing, transparency, and its changed position on an opt-out exception. (UK Government, 2026)

AI, copyright and the creative industries: A parliamentary inquiry examines copyright enforcement, training-data transparency, and licensing options for creative industries. (House of Lords Communications and Digital Committee, 2026)

Children and young people's reading in 2026: The latest annual survey places concerns about future readers alongside evidence on reading enjoyment and frequency. (National Literacy Trust, 2026)

Research findings are available online in The Impact of Generative AI on the Novel.

The original story "Half of surveyed UK novelists fear AI could replace their work entirely" is published in The Brighter Side of News.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.