Literary agents are rewriting their submission policies in response to a surge of manuscripts and cover letters produced with artificial intelligence, signaling a growing backlash within the publishing world against the technology’s encroachment on creative authorship. Industry professionals say the volume of AI-assisted submissions has risen sharply over the past year, forcing agencies to clarify that only fully human-created work will be considered.
Executives at Greene & Heaton reported a noticeable shift in the nature of incoming manuscripts, with managing director Antony Topping describing widespread changes in style and presentation that suggested the use of generative tools. The agency now explicitly states that submissions must be entirely original and warns writers that even seemingly minor AI editing can “flatten” their voice. Its updated guidance rejects any use of AI in manuscripts, proposals, synopses, or cover letters.
Other agencies are adopting similar measures. Representatives at Bright Agency acknowledged that illustrators may sometimes rely on AI to help articulate written materials, but stressed that the same allowance cannot be extended to authors, arguing that algorithmic assistance risks diluting the individuality that publishers seek to champion.
At Eve White Literary Agency, new submission guidelines emphasize that all entries are reviewed by experienced professionals without technological mediation and that any writing generated through AI will not be accepted. The agency has framed the policy as a defense of authenticity in storytelling at a time when automation is becoming increasingly sophisticated.
According to reporting by The Bookseller, agencies across the sector have observed a marked uptick in AI-generated queries and manuscripts, creating practical and ethical dilemmas about evaluation, originality, and disclosure. Firms such as Janklow & Nesbit say distinguishing between human and machine-produced content is becoming harder as the technology improves, presenting what they describe as a significant long-term challenge for the industry.
The debate is unfolding alongside mounting legal disputes over how AI systems are trained. Last year, AI developer Anthropic agreed to a £1.1 billion settlement in a class-action lawsuit brought by writers who alleged their copyrighted works had been used without permission to train language models. Similar claims have been filed against major technology companies including OpenAI, Meta, and Microsoft, underscoring the widening clash between Silicon Valley and creative industries.
Trade bodies such as the Publishers Association are lobbying governments to ensure that copyrighted material cannot be used for AI training without consent or compensation, reflecting broader concerns across film, music, journalism, and publishing.
Agents say their stricter submission standards are less about resisting innovation and more about preserving the distinct human perspective that defines literature. Yet with generative AI tools becoming more advanced and accessible, many acknowledge that enforcing these boundaries may become increasingly difficult, setting up a prolonged struggle over creativity, ownership, and the future of authorship itself.
