British novelists are confronting an existential threat from artificial intelligence, with half fearing their profession could become obsolete according to groundbreaking research from the University of Cambridge.
The Anxiety Behind the Statistics
The University of Cambridge's Minderoo Centre for Technology and Democracy conducted comprehensive research among 258 published novelists and 74 industry professionals to gauge attitudes toward AI in British fiction. The findings, released on Thursday 20 November 2025, reveal profound concerns about the technology's rapid advancement.
Exactly 51% of surveyed authors expressed fear that AI systems could eventually replace human novelists entirely. This anxiety persists despite most writers avoiding the technology in their own creative processes.
More immediately pressing are financial worries, with 85% of novelists believing their future earnings will suffer due to AI competition. Already, 39% report their finances have taken a measurable hit from AI's incursion into publishing.
Industry Voices and Economic Realities
Bestselling author Tracy Chevalier, renowned for Girl With A Pearl Earring and The Glassmaker, articulated concerns shared by many in the literary community.
"I worry that a book industry driven mainly by profit will be tempted to use AI more and more to generate books," Chevalier stated in her survey response. She drew a compelling comparison to consumer behaviour in other industries, noting that cheaper machine-made products often displace handcrafted alternatives in the marketplace.
The financial vulnerability of professional writers amplifies these concerns. The research highlights that the median annual income for novelists stands at just £7,000, forcing many to supplement their earnings through related work like audiobook narration, copywriting, and ghost-writing.
Copyright Infringement and Unauthorised Training
Perhaps the most contentious issue identified in the study involves copyright and the unauthorized use of authors' work to train AI systems.
59% of novelists surveyed reported knowledge that their copyrighted material had been used to train AI models without permission. Among these authors, 99% confirmed they never granted consent for this use, and 100% stated they received no compensation for the appropriation of their intellectual property.
The report references a significant legal settlement earlier this year where AI company Anthropic agreed to pay authors $1.5 billion (£1.2 billion) to resolve claims it had downloaded over seven million pirated digital books for training purposes. However, the court stopped short of ruling this activity as copyright infringement, comparing it instead to human reading for inspiration.
Practical Impacts and Government Response
Beyond theoretical concerns, authors report concrete negative effects already occurring. Respondents described discovering "rip-off AI-generated imitations" of their original works and even books published under their names that they didn't write.
Last year, the Authors Guild warned about a surge of low-quality sham 'books' on Amazon driven by AI accessibility, prompting the retail giant to limit daily publications on its Kindle self-publishing platform.
The report's authors urge government intervention, specifically recommending strengthened copyright protections to safeguard creative industries. Professor Gina Neff, executive director of the Minderoo Centre, criticised proposed copyright exceptions for text and data mining, arguing this approach "prioritises access to data for the world's technology companies at the cost to the UK's own creative industries."
A government spokesperson responded that they are committed to "working with both the creative industries and AI sector to drive AI innovation and ensure robust protections for creators."
Despite the overwhelming concerns, the research did identify some authors finding productive uses for AI tools. Novelist Lizbeth Crawford reported using AI as a writing partner to identify plot holes and streamline editing, potentially increasing her annual output from one novel to three, with ambitions to reach five.
Nevertheless, 67% of novelists never use AI for creative work, reflecting the deep reservations within the literary community about embracing technology they perceive as threatening their livelihoods and artistic integrity.
