Prize-winning author defends using ChatGPT for her novel
Japanese novelist Rie Qudan, 34, has sparked considerable debate with her latest work, Sympathy Tower Tokyo, a novel that recently clinched the prestigious Akutagawa prize. The controversy stems not solely from its compelling narrative but from Qudan’s candid admission that parts of the book were generated using ChatGPT.
Speaking from her home near Tokyo, ahead of the novel’s English translation release, Qudan addresses the unease some feel about AI’s role in creative endeavors. She expresses no distress over her work potentially training AI models, asserting a unique, uncopyable essence within her writing. The author clarified that the approximately 5% of the novel attributed to AI comprised direct exchanges between a character and ChatGPT, serving as a narrative device. More broadly, Qudan revealed that interactions with AI proved a significant source of inspiration, offering fascinating insights into human thought processes. Her use of AI, therefore, aimed not to deceive, but to illuminate its potential effects and reflections of human cognition.
Sympathy Tower Tokyo centers on Sara Machina, a Japanese architect tasked with designing a new high-rise to house convicted criminals, conceptualized as a beacon of “compassionate comfort”—a notion one character ironically labels as reflecting “the extraordinary broadmindedness of the Japanese people.” Sara, herself a victim of violent crime, grapples with the appropriateness of such a sympathetic approach to offenders. This central theme was partly inspired by the 2022 assassination of former Prime Minister Shinzo Abe, an event that saw the shooter’s background elicit unexpected public sympathy in Japan due to his difficult religious upbringing. The novel probes public attitudes towards criminals through a satirical “Sympathy Test” for potential tower residents, with AI ultimately deciding who is deemed worthy of compassion.
Despite the Akutagawa prize bringing a sense of liberation after a previous nomination in 2022, the novel’s AI component has drawn significant attention. Yet, Qudan observes that the ensuing discussion among Japanese readers has focused more on the language itself and how linguistic shifts over recent decades influence perception and behavior. Indeed, language forms the bedrock of Sympathy Tower Tokyo, where words are not merely tools of expression but profound revelations of reality. As one character states, “Words determine our reality.”
A key linguistic debate in the novel revolves around the increasing prevalence of katakana, the Japanese script primarily used for foreign-derived words. Unlike traditional kanji characters or hiragana script, katakana words often sound milder and more euphemistic, potentially allowing for the avoidance of discriminatory phrasing. Characters like Sara lament this trend, viewing it as an abandonment of traditional Japanese language, while others resist what they see as a “wretched proliferation.” Qudan, born in 1990, notes that for her generation, katakana has become an unquestioned standard.
This linguistic evolution carries urgent political implications. Qudan connects it to the recent surge in support for the far-right Sanseito party, which, campaigning on a “Japanese people first” slogan reminiscent of Donald Trump’s “America first,” saw its representation in the upper house of parliament grow from one to fourteen seats. Qudan explains that Sanseito deliberately employs the katakana equivalent for “first” in its slogan. This linguistic choice, she argues, replaces potentially negative associations with neutral ones, effectively softening the message and preventing immediate negative reactions. It creates a form of plausible deniability, a calculated move that demands scrutiny. “When someone uses katakana,” Qudan concludes, “we should ask: what are they trying to hide?” This question underscores her novel’s profound inquiry into how language can both conceal and reveal the complex realities of Japanese society, including its ongoing struggle with diversity and underlying prejudices.