Select Language

English

Down Icon

Select Country

Spain

Down Icon

Books vs. Artificial Intelligence: Turning the Page?

Books vs. Artificial Intelligence: Turning the Page?

Recently, writer Isabel Allende recounted how her son, to demonstrate to her how writing would no longer make sense in the future, asked an artificial intelligence program to write a story about a shy boy whose dog saved him from bullying . The result, according to Allende, was "almost identical" to her book, *Perla, la superperrita*, although perhaps a little "flatter." Although Allende spoke enthusiastically about the experiment, it's not hard to imagine how that same prospect could be troubling to many other writers and publishing industry players.

According to a survey conducted by Proyecto451, an agency that provides digital services to publishing companies, while among more experienced editors a positive view prevails regarding the impact of AI on the publishing world, for translators, authors, and especially illustrators, the relationship is reversed.

With any technological revolution, fears of job replacement are nothing new. But what's perhaps most disturbing in the case of AI is its ability to generate pieces that, depending on how you look at it, could be considered artistic—something that until now was thought to be the exclusive domain of humans .

This gives rise to a range of philosophical debates about what a creative act truly entails, in addition to the legal challenges and specific problems surrounding the alteration of forms of production and circulation.

In 2023, more than 15,000 authors from the US organization The Authors Guild —including Margaret Atwood and Jonathan Franzen —signed an open letter demanding that companies like OpenAI and Meta stop using their works without permission or compensation. “These technologies mimic and regurgitate our language, stories, styles, and ideas. Millions of copyrighted books, articles, essays, and poems are 'fed' to AI systems,” they noted in the letter.

According to an investigation by The Atlantic , Meta used Library Genesis , one of the largest repositories of pirated books circulating online, as a data set, thereby violating the copyright of thousands of writers.

Margaret Atwood holds an Extinction Rebellion sticker before an interview for the Writers Rebel podcast in London, Britain, October 14, 2019. Extinction Rebellion/Writers Rebel Podcast Photo: Reuters Margaret Atwood holds an Extinction Rebellion sticker before an interview for the Writers Rebel podcast in London, Britain, October 14, 2019. Extinction Rebellion/Writers Rebel Podcast Photo: Reuters

Currently, copyright regulations, based on traditional understandings of authorship and creativity as a product of human invention, do not offer sufficient responses to this new scenario. Faced with this gap and in line with organizations from various countries, the Union of Writers of Argentina released a statement signed by more than 150 authors in April demanding that publishing contracts include "clauses granting creators the power to restrict, limit, or even prevent the use of AI in their works and/or for AI to be used in training such technologies." For the time being, in the absence of general legislation, the solution presented is private agreements.

The claim, in addition to defending writers' already frequently undervalued incomes, aims to uphold the quality dimension. According to the Authors Guild letter, the market could be flooded with "mediocre" books. In fact, Amazon was forced to limit self-publishing after an avalanche of AI-written books used the service.

The proliferation of pseudo-books opens the door to scams: books that promise to be one thing from their digital covers but then have content that doesn't match, authors with fake identities, and faces generated by computer programs. Writer and editor Jane Friedman took Amazon to court after accidentally discovering that the platform listed books with her name that she hadn't written. Beyond her personal case, the author also noted with concern another phenomenon: some people, based on short synopses of advance publications, were generating books before they were published.

The prospect that years of work could be skipped by simply writing a couple of prompts could, for example, displace the arduous work done in writing workshops. However, once again, will it be enough to ask AI to generate ten possible alternatives for the same text?

While its uses for speeding up workflows might be more readily welcomed by editors and authors—for example, to detect spelling or formatting errors (or even as a tool for rewording sentences)—its usefulness in defining the quality of a text is more ambiguous.

AI learns from the past, drawing on what has already been produced to create. In principle, this wouldn't be very different from the process of a human being being inspired by their ancestors. However, while real-life creators take responsibility for their word (or can be held accountable), it's unclear how that notion translates to the objects produced by AI.

The so-called "bias problem" aims to highlight the social and cultural imprint underlying these systems, as they can reproduce and propagate pre-existing prejudices depending on the data they crunch. In this case , AI could generate text advocating torture or inciting suicide purely mechanically, without any ethical consideration. Faced with this problem, companies have had to take measures to counteract the harmful effects, introducing restrictions and refining databases.

But art often walks the ledge of limits. The representation of violence in the context of a work, for example, is reinterpreted, taking on new dimensions that might escape the literal interpretation of an AI. Could a restricted AI with moral guidelines write Nabokov's Lolita , or would it respond by refusing the prompt asking it to adopt the perspective of that despicable narrator?

Dressed in red and with pigtails. Monster. Dressed in red and with pigtails. Monster.

Recently, illustrator María Verónica Ramírez discovered that an AI-generated reproduction of her work, created in support of the Garrahan Hospital , had gone viral on social media. However, the replica had transformed the image into a linear representation, stripped of the metaphor present in the original drawing and removed from the universe of meanings of her character, Monstriña .

“The best literary translations offer more than simple accuracy, more than literal fidelity to the words that make up the sentences,” said translator Polly Barton when asked by The Guardian about the launch of GlobeScribe , a literary translation service that, its creators promise, produces results indistinguishable from human translations. For Polly Barton, as for other colleagues, the true value of a translation lies in the interpretation of a source context, in the search for the reproduction of rhythms or atmospheres that literary texts produce in readers .

The replacement of these jobs, while potentially possible, will not occur without resistance. In Spain , for example, several bookstores decided to remove the book Joan of Arc , published by Planeta , from their shelves when a renowned illustrator published a message demonstrating that the cover had been designed by an AI .

An artificial intelligence-powered robot called Ameca, developed by Engineered Arts, demonstrates imitations in London. Photo: EFE/Tolga Akmen An artificial intelligence-powered robot called Ameca, developed by Engineered Arts, demonstrates imitations in London. Photo: EFE/Tolga Akmen

Perhaps this act of solidarity between two actors in the book ecosystem could also be explained by a sense of shared threat. If a bookseller's profession could be evaluated by their ability, among other things, to offer customers recommendations, platforms like Goodreads or The StoryGraph behave like artificial imitators that generate tailored suggestions.

The loss of complexity in favor of transparency and the homogenization of taste through the algorithmic repetition of the same things seem to foster a climate of cultural flattening. In addition, there are studies that demonstrate the negative impact on people's cognitive ability due to the constant use of programs like ChatGPT , although this could well be reminiscent of Plato's complaints about the written medium.

Despite all the alarms, there are those, like Isabel Allende, who see the development of AI as both a challenge and an opportunity for experimentation. Spanish writer and literary critic Jorge Carrión published Magnetic Fields , in which he credited two AI systems as co-authors. It was a tribute to the work of surrealists André Breton and Philippe Soupault , in which he extended the idea of automatic writing to a dialogue between two machines.

"If for years we grew accustomed to the spell checker in our word processor, now we are the spell checkers and editors of the texts created by OpenAI's program ," Carrión stated in an article in La Vanguardia. How much truth there is in this diagnosis will be proven in the years to come.

Clarin

Clarin

Similar News

All News
Animated ArrowAnimated ArrowAnimated Arrow