Frankentext: workshop brings together six Italian writers with AI

© 2025 EPFL

© 2025 EPFL

In May 2025, Italian writer Caterina Serra brought a group of Italian writers to EPFL for a weeklong workshop that aimed at creating a literary work using generative AI.

The writers – Daniela Cascella, Fumettibrutti, Lorenzo Iervolino, Djarah Kan, Federico Zappino, and Caterina Serra – spent one week together at EPFL and the Fondation Jan Michalski exploring the different kinds of generative AI tools, creating prompts, and writing collaboratively with Tammara Leites, an artist, developer and interaction designer working on projects exploring the creative use of AI.

Serra conceived and coordinated this project within the EPFL College of Humanities, following the lead of director Frédéric Kaplan, who was keen to work at the intersection of literature and artificial intelligence. The workshop was an immersive experience lasting one week, allowing the six writers to engage with the potential and challenges of using artificial intelligence and to discuss the impact that interaction with digital technology has on the creative process, on reality, and on the worlds they write about and within.

“I took inspiration from the Villa Diodati literary experiment of 1816,” says Serra, referring to the period when Lord Byron, Dr. John Polidori, Mary Shelley, Percy Bysshe Shelley, and Claire Clairmont stayed at the Villa telling each other horror stories, eventually leading to the publications of Frankenstein by Mary Shelley and The Vampyre by Polidori.

Serra explains: “The idea was to have a real, concrete, and material experience with something that is not apparently physical, like artificial intelligence. I thought that it was important for us, now in this age, to have the real experience of writing in an embodied way. I chose the other five writers, each of whom has very different styles, ages, identities. Maybe the only thing we have in common is the idea that writing arises from looking at reality. It is a political act, I would say.”

The workshop was conducted in Italian to serve as a monolingual model adaptable to other linguistic contexts for future expansion. Initially, the idea was that each writer would work on their own piece, culminating in a hybrid collection exploring the relationship between writing and technology. However, once the workshop began, the group instead decided to work together, spending nearly a half hour collaborating on a prompt to give the generative AI program they chose to use, Chat GPT.

“Working on the prompt is when things started to become more interesting for us,” said Cascella at the conference the group held at the end of the week. After the week, she adds that “I feel a lot more confident about what I can do as a writer”.

“At first I thought that AI could take my job, but after working with it, it gave me more confidence in my own writing, because AI cannot do what I can do,” added Kan.

Indeed, the rest of the group echoed this opinion that while aspects of the generative AI program were impressive, they did not see it having the potential to create literature or critical thought pieces. Fumettibrutti, who is a graphic novelist, was especially unimpressed, as ChatGPT censored her work and was unable to produce anything that remotely resembled it.

Before the workshop, none of the writers had worked with generative AI before. Serra, to prepare herself to lead the group, used some different LLM programs, read books and articles about AI as well as the related economics and politics.

While Serra also found the content produced by Chat GPT to be lacking in many ways, she found working with the generative AI to be “seductive, as a product of a capitalistic form of addiction.” The group sometimes called the machine “cara" (sweetie) or “amore” (love). And because the group was working in Italian, they used gendered pronouns – he or she – to refer to the program, as if it were a person, anthropomorphising the machine.

“We discussed a lot about the ‘humanity’ of the AI because we expect something human to come from the machine. But that’s a mistake; it’s our bias, our perhaps too naive approach of embracing technology as something to be accepted as it comes to us. If a chatbot tells a story about the world, with what gaze is it doing so? With what ethics, with what care? My idea of literature comes from emotions, goes through critical thinking, freedom of expression, often taking a long time that is also made up of non-writing time. It is not a product, but an act, often an act of revolution.”


Author: Stephanie Parker

Source: College of humanities | CDH

This content is distributed under a Creative Commons CC BY-SA 4.0 license. You may freely reproduce the text, videos and images it contains, provided that you indicate the author’s name and place no restrictions on the subsequent use of the content. If you would like to reproduce an illustration that does not contain the CC BY-SA notice, you must obtain approval from the author.