Franz Kafka wrote a short story entitled Report to an Academy, in which an ape gains human intelligence and gives a lecture about its past as a wild animal. Something similar happens when the signer asks ChatGPT about their own limitations: “Can you write a good poem?” The language model – as these artificial intelligence tools are called – which will stop at nothing, answers: “Although it is possible to write poetry without a human author, most poetry that is considered relevant and meaningful is usually written by one.” Poet co-authors a distinctive identity and voice. He admits with impossible aplomb that the work of an author without a biography does not usually arouse much interest among readers.
More information
ChatGPT may be the most prolific poet in the history of literature since it launched just a few months ago. Thanks to huge deep learning neural networks, they are able to write texts in a few seconds, based on very simple instructions and without plagiarizing a single sentence from the Internet. However, he cannot be considered the legal author of any of his works. “Perhaps a machine can create art, but it is not protected by intellectual property, because for that there must be an author, and as such, except in rare cases, it can only be natural persons,” says Mario Sol Muntañola, lawyer, intellectuals expert Property.
Guillermo Marco and Julio Gonzalo, UNED researchers in the field of natural language processing, have spent years studying the limitations and possibilities of language models like ChatGPT, and they say they are skeptical about “the hype” surrounding the technology. “These models work like networks of artificial neurons, which are like our brains, and they learn to write by cognitively simulating our reading,” explains Marco, also author of the poetry collection Other Clouds (Rialp, 2019 ), for which he wrote a second in 2019 Adonáis Poetry Prize.
In one of their first experiments, Marco and Gonzalo asked a group of participants to rate six aspects of the synopses of books and films that this artificial intelligence produced. “We gave it a made-up name, and the machine made an argument for that title,” explains Gonzalo. The result was that the language model performed better than humans in all areas except creativity. The researchers decided to focus on this specific aspect and changed their subject of study. “We started experimenting with poetry because synopses, at least for humans, are not inherently creative texts,” they say. Before measuring the results, they suggested extracting a definition of what users understand by creativity. “It’s different for everyone, but almost everyone agreed that for them creativity was the original, the unusual, something they had never seen before,” Marco explains.
The Frenchman Philippe Soupault is considered one of the fathers of automatic poetry. Marc GANTIER (Gamma-Rapho via Getty Images)
The duo concluded that these language models are not designed to be particularly creative. “Before they are made to imitate rather than be original. They learn to say the least surprising,” they affirm. Marco explains it with an example: “If you give him the sequence: ‘Heaven is…’ and you ask him to complete the sentence, he will always choose the blue word from the 50,000 words he has learned. because that’s most likely”. ChatGPT’s priority, they explain, is to preserve the meaning of the text, making it difficult for a sentence to be aesthetically striking. “Then it turns out there are people who the Machine asking questions that already have such a high originality component that they have no choice but to improvise,” clarifies Julio Gonzalo. “I read an example where someone asked ChatGPT to write a Bible-style story about a person writing who put a butter sandwich in the hole of a VHS video player. The result was hilarious.”
It’s also important to consider that ChatGPT doesn’t understand the words it’s learning. “All your knowledge is intuitive. By reading they learn what language is. But they don’t have the capacity for reflection or rational thinking. Precisely because of this, they learn to imitate sonnets, but are unable to explicitly see that there is a rule relating a certain type of rhyme to a certain number of verses,” comments Gonzalo.
Marco emphasizes that these language models will always be limited to the input sequence, i.e. to the instructions of a human. “You will never have an idea for the work. You will never have intention. That intention will always come from a human being in the moment. What would be truly artistic is if he refused to write or chose to write it his own way.”
surreal neurons
In the book Non-creative Writing: Managing Language in the Digital Age (Black Box, 2015), writer Kenneth Goldsmith writes: “Perhaps the great writers of the future will be those who write the best programs for manipulating, analyzing, and disseminating language can practice language”. It refers to the possibility of literature becoming a collaboration between machines and humans that seeks not to replace human creativity but to augment it.
Writer Jorge Carrión has just published Electromagnetic Fields. Theories and Practices of Artificial Writing (Black Box), a book he wrote with the help of Taller Estampa’s engineers using a GPT-2 and 3. His starting point is The Magnetic Fields by André Breton and Philippe Soupault, which is considered the birth of automatic poetry in 1919. In his book he compares the arrival and influence of the Surrealist group on literature and culture in the first decades of the 20th century, with the expansion of models of language: “If the transition between conscious and unconscious writing marked these years, the writing that produced by machine learning and other forms of artificial intelligence gives ours a special vibration.”
Speaking to ICON, he assures that “algorithms write very well, they write almost perfectly, they access areas that are taboo for humans, but they are still incapable of brilliance, of metaphor, of knowledge of the best poetry. That doesn’t mean they don’t write better, as many poets influence the poet through experience and evolution.
However, in the introductory chapter of his book, he does not hesitate to say that the arrival of a technology capable of writing good literature is only a matter of time: “The intelligence of algorithms, robots, neural networks or programs or formulas , the artificial intelligence that does not yet exist and is therefore nameless, will ultimately be capable of metaphor and irony, in new, cinematic and undoubtedly literary forms.
You can follow ICON on Facebook, TwitterInstagram or subscribe to the newsletter here.