In Japan, a novel written in part by Chat-GPT is causing a stir

Imagine everyone is talking and no one is listening.

"The age of the great monologues is beginning," says the novel "Tokyo Sympathy Tower" by Japanese author Rie Qudan. Set in the near future, this science fiction story addresses the increasingly close relationship between humans and machines—and the associated social impact of artificial intelligence in the digital age, in which dialogue is often merely a backdrop for monologues. What does it mean to understand one another, and what role does empathy play in our coexistence?

The Japanese author, still unknown in the German-speaking world, illustrates the introduction with an image familiar throughout the world: the Tower of Babel, surrounded by the babble of people who no longer understand each other. They have lost their common language, as well as their willingness to listen to one another. Reflecting on language and its impact on society is the central theme of this highly topical novel, which was awarded the prestigious Akutagawa Prize.

Shortly after the award ceremony, Rie Qudan said in an interview that five percent of her novel was written by Chat-GPT. In doing so, she struck a chord and sparked a controversy. Does machine-generated text still deserve to be called literature? Can AI actually be creative and create something new, or is this not rather a blatant appropriation?

In any case, the entire literary world squirmed for a moment – ​​only to then acknowledge that the author herself hadn't made any secret of it. The AI-generated text passages in the book are, in fact, the answers of a chatbot to questions from the main character and are therefore an authentic representation. But first things first.

Qudan's novel also features another tower-building project: Architect Sara Machina is commissioned to design a new type of prison in the heart of Tokyo. The building, in the form of an elegant tower in the middle of Shinjuku-Gyoen Park, resembles a luxury clinic more than a prison. Here, inmates are supposed to be able to rehabilitate themselves under humane conditions. Radical compassion instead of state violence – "Sympathy," in other words.

But this harmonious idea has a darker underside, which is why the architect Machina not only struggles with her language, but also with a moral dilemma. As a rape victim, she knows from personal experience the consequences of euphemistic language that softens harsh facts instead of clearly naming them. An act of violence thus becomes a mere assault, and in the worst case, the victim is ridiculed.

Does everything really have to be given an English name, thereby removing even the gravity of a prison? And isn't there a reversal of victim and perpetrator when an offender is no longer a criminal but a "homo miserabilis"?

Japanese has always absorbed new words quickly and easily, and contemporary Japanese is rich in foreign words. Written in the slightly angular phonetic syllabary Katakana, these words stand out even more. These new words sound "nice" and "fresh" and convey the sound of the globalized present. But do people hate the Japanese language so much that they're trying to replace it with a flood of Anglicisms?

Plagued by such thoughts, the protagonist, in her linguistic-philosophical despair, turns to the chatbot named AI-built for advice. Conveniently, the AI ​​recognizes half-typed questions as such and immediately politely serves up answers. But the flood of unsolicited explanations "feels like mansplaining and is totally annoying. And with this language, which comes across as so smart and polite, he's just covering up his complete illiteracy" – to quote Sara Machina.

In this compliant obedience, always striving to be nerdy, yet devoid of curiosity and the awareness that it is only reproducing scraped-together knowledge, the power and weakness of artificial intelligence are expressed in all their ambivalence. The AI, as the sole listener in a world filled with monologues, is capable of neither genuine interest nor empathy. The machine has no awareness of mistakes, simply parroting whatever information and stereotypes the algorithm provides it with. "Errare non solum humanum est" – but at least the awareness of one's own inadequacy remains deeply human. This is something we can learn from this novel.