The emergence of generative artificial intelligence has transformed the literary landscape in just a few years. What once seemed like science fiction is now a reality affecting writers, editors, and readers. The question is no longer whether AI can write, but what it means to be an author when a machine is involved in the creative process.
The debate has intensified with studies revealing a paradox: readers prefer the clarity of AI-generated texts, but at the same time, they value the work more when they believe it is human-generated. This contradiction reflects the complexity of a relationship that still lacks clear rules.
The reader's paradox
Recent research, such as that conducted by Villanova University, indicates that readers can barely distinguish between stories written by humans and those written by AI , but they give higher ratings to those they believe to be human-authored. This bias is explained by prejudices against technology, even though many people privately enjoy machine-generated works.
Educator José Manuel Suárez Sandomingo, president of APEGA, argues that the main advantage of AI-generated texts is their clarity: "People gravitate towards what's fast, what's agile, what's easiest to read." However, this preference clashes with the distrust expressed by many readers on social media, where they assert they would never buy a book written by a machine.
The contradiction is resolved when the context is analyzed: AI produces efficient texts, but lacks the emotion and sociocultural context that characterize human writing. As the writer María Oruña points out, "it is absolutely impossible for AI to make the same mistakes and blunders I do, or to have the same narrative approaches as I do."
Writers at the Crossroads
The literary world is divided. Some authors, like the physicist and writer Agustín Fernández Mallo, use AI as a research tool, but warn of its dangers: "AI gives you everything ready-made; you don't develop your own techniques and methods, so your brain ends up stagnating." Others, however, reject it completely, believing that handing over writing to a machine means losing the essence of the craft.
Between these two extremes lies a gray area. Many writers use AI to revise, structure, or overcome writer's block , but they draw a line when it comes to literary creation. Writer Miguel Ángel Hernández has argued that AI models tend to homogenize language, which could eliminate stylistic diversity and the subtle pleasures that only writers appreciate.
The concern isn't merely aesthetic. A survey by the Australian Society of Authors revealed that 87% of writers have not authorized the use of their works to train AI systems, and 89% fear that these tools will copy their style. This concern has led to class-action lawsuits against major tech companies.
The Anthropic scandal
The most notorious case is that of Anthropic, the company behind the Claude chatbot. Declassified court documents revealed Project Panama , a secret operation to buy, scan, and destroy millions of physical books in order to train its models. The company acquired copies from secondhand bookstores, cut off their spines, and digitized the pages before sending them for recycling.
The strategy aimed to obtain high-quality texts, edited and reviewed by humans, to avoid the "model collapse" that occurs when training with AI-generated content. However, the operation also included downloading more than seven million pirated books from repositories such as LibGen and Pirate Library Mirror.
The case ended in September 2025 with a $1.500 billion settlement, the largest copyright award in history. The court ruled that the use of legally acquired books was protected under the fair use doctrine, but not pirated material. The ruling sets a precedent for other companies like OpenAI, Google, and Meta , which face similar lawsuits.
The future of authorship
The publishing industry still lacks clear answers. Should readers be informed when a book has been created with AI assistance? To what extent can a machine intervene before the work ceases to be human-made? These questions divide authors, publishers, and readers.
Meanwhile, technology advances and becomes increasingly difficult to detect. Readers may say they will never read a book written by AI, but perhaps they already are without realizing it . The paradox is that, although AI mimics language, it lacks the memory and finiteness that give meaning to human creation.
The writer Agustín Fernández Mallo sums it up with a metaphor: AI is a fallen angel, an entity with its own will that we have created, but which has neither a body nor scars. Therefore, although it can generate perfect texts, it will never be able to transcend as human literature does. The final decision rests with the readers and publishers, who must define the limits of this new tool.
