Can an AI write your next novel? Ethics and Creativity in Literary AI
I see this issue from inside the industry: AI generates textures, ideas and drafts, but the human spark is still the real fuel. In 2025, Spain is in a tug-of-war between tools like Jasper, Sudowrite and Writesonic and the craft. The data is not anecdotal: a study by the Autonomous University of Madrid in February 2025 puts 80% of Spanish writers using AI at some point in the process. This is not a fad, it is a work habit.
The editorial also changes. Newtral verified that, in 2023 and 2024, 20% of manuscripts received were AI-generated and 40% showed signs of AI-assisted plagiarism. This is not a minor detail for anyone who wants to publish rigorously.
The use of AI in the creation of literary texts grows further from 85% in two years. And we are not just talking about experiments, but about practices that cross editorial thresholds.
In this context, the underlying question is not whether AI can write a novel, but whether AI can write a novel worth reading. This distinction, which sounds like academic jargon to many, is everyday practice for authors, agents and publishers. If the novel comes out pitch-perfect and flawless, what role does the human author when the algorithm already knows what pace the reader wants, what cliffhanger works and what sentence leaves a mark? That's where the discussion gets serious.
My experience from years of working with authors indicates that AI is a powerful tool to speed up the draft, to explore plots and to generate variants. Jasper allows for text generation, brainstorming and multi-voice branding modes. Sudowrite helps with creativity and proposes twists in plots and characters, Writesonic offers an editor with real-time AI and multilingual generation.
But using these tools without judgement means getting lost. AI has no market intuition, does not understand the pulse of a community of readers, nor does it know when a scene needs silence, pain, or restrained ambition.
The ethical boundary is real. In 2024, Spanish publishers suspended the reception of texts in the face of an avalanche of AI-generated works. It wasn't just fear of quality, it was fear of intellectual property and authenticity. ANIM's XVII Cibercertamen Literario requires a declaration of whether AI was used. This is already the norm to avoid false authorship and ensure transparency. In my opinion, this clarity is the basis for a sustainable publishing economy: if someone signs an AI-generated work without acknowledgement, the whole ecosystem is undermined, from distribution to sales.
For the creator, the key is to combine. AI to cut research time, to suggest structures, to check for consistency, but always with the human stamp. An AI-generated draft may contain logical gaps or stylistic traps; the author must detect and resolve them, and provide the tone, voice and look that makes a novel unique.
Modern literary criticism reminds us that poetry and deep narrative are still resistances of meaning that machines have not learned to fully flaunt. Jessica Mariana García Vázquez points out that AI can imitate styles and generate coherent verse (but the experience of deep meaning remains human territory).
The historical context helps to understand the evolution. Between 2016 and 2021, GPT-2 and GPT-3 opened up automatic generation, but with consistency limits. Then came ChatGPT and multimodal models, capable of learning from the user and imitating styles with speed. In this framework, AI does not destroy the craft, it transforms it: it changes the way a work is planned, edited and presented, yes, but it does not eliminate the need for a critical eye, a market vision and a commitment to quality.
So what do you do as an author or editor who wants to maintain integrity? First, be clear with the reader and with the team. State whether AI was involved in the creation and what roles it played, not hide it behind a veil of «creative collaboration». Second, use AI as an ally, not a substitute. In practice, I can use AI to generate plot ideas, check for inconsistencies, propose changes in pacing, and translate into other languages to assess scope. But the narrative voice, emotional arc, characterisation and ending must be the product of human experience.
If AI-generated passages are used, they must be rigorously checked and registered in terms of copyright. AI-assisted etiquette is the basis of trust with publishers, agencies and readers.
The battle between creativity and tool is not one of antagonists, it is one of partners. Successful authors who incorporate AI judiciously establish processes: AI-driven ideation, AI-guided writing to maintain consistency of style, and strict revision by humans. In my experience, that flow generates speed without sacrificing quality, and avoids stagnation of narrative voice. It's also worth thinking about distribution and marketing from the start. If AI helps create content, the editing of the proposal and the creation of the cover should remain in the author's control so as not to lose brand identity.
In 2025, the institutional conversation about AI and literature is more serious: the goal is not to ban AI, but to regulate it so that it contributes without eroding authorship and originality. Conscious use of AI, accompanied by rigorous human review, can speed up the process, broaden the scope and maintain literary quality.
For those who want to publish without relying on a large traditional publisher, AI can be a decisive resource, as long as the final work has a clear human origin. At Infinity Ideas Studio, we have seen cases where the combination of self-publishing and co-publishing benefits from AI for repetitive or idea-generating areas, but the final authorship and responsibility remains in the hands of the author and the editorial team. Our approach is transparency, control and autonomy for the creator to publish with clarity, professionalism and real results.
If you're thinking about your novel, here's my practical advice. Define what role AI wants to play in your process: ideation, structure, style, revision or market. Set clear boundaries: which parts are 100% human and which parts are assisted. Document the use of AI for each manuscript.
Review the generated sections with a critical eye and rewrite them if they don't fit your voice. Keep a rights register for each piece that uses AI and be prepared to explain your process to publishers and competitions.
In the end, yes, an algorithm can write a novel. But the novel that reaches the reader is the one that passes through the human filter before, during and after the machine's intervention. That combination, when done with ethics, clarity and rigour, offers clarity and an authentic voice that the reader can recognise. And therein lies the difference between an AI-generated utilitarian work and a novel that endures because someone put the experience, the courage and the eye on every page.
Best AI 2026 for Writing a Book
Clarifications «Mi alma grita lo que el hombre calla» - Book by Isabel Covada.
Copyright and self-publishing - clear guidance and effective protection
Español 
English (UK)