Semafor Signals
Supported by
Japanese author used ChatGPT to write an award-winning novel
Insights from the New York Times, Sixth Tone, and The Times
The News
A Japanese author disclosed she used generative AI to write an award-winning novel, demonstrating the technology’s growing impact on culture. Rie Kudan, 33, announced at the Akutagawa Prize ceremony this week that about 5% of her book Tokyo-to Dojo-to (“Sympathy Tower Tokyo”) included verbatim sentences generated by ChatGPT.
Last year, a Chinese professor also used AI to write a science fiction novel in just three hours, and went on to win a national competition. Kudan said she turned to AI to help her find “soft and fuzzy words” that embodied the muddled themes about justice present throughout her book, which the prize committee called “flawless.”
“I want to use the words with care, and to think about the positive and negative aspects of language,” Kudan said. The judges determined her book was a “highly entertaining and interesting work that prompts debate about how to consider it.” The use of generative artificial intelligence has become hotly debated among creative professionals, many of whom fear rapid adoption of the technology could pose a threat to their livelihoods.
SIGNALS
Authors have already begun using AI in their writing
A Tsinghua University professor won a national science fiction honor after generating his entire book, Land of Memories, using AI technologies. The South China Morning Post reported that not all the judges were able to tell the difference: One said the language was “weak,” but another argued she paid attention to “creativity and scene description.” A book critic for The New York Times argued that AI won’t completely replace humans in literature, stating that authors are embracing “the latest iteration of an ancient literary conceit: the fantasy of a co-author, a confidant, a muse.” “Poets and novelists once turned to séances, Ouija boards and automatic writing for inspiration. Now they can summon a chatbot to their laptops,” he added.
Countries differ in their legal approaches to AI-generated art
A Beijing court recently granted copyright protection to a picture generated using the AI text-to-image generator Stable Diffusion, stating that it was “directly created from the plaintiff’s intellectual input,” and “manifested in (his) individual expression.” One copyright expert told Sixth Tone that the case opened the door for other “half-human, half-AI works” to potentially be given intellectual property protections. In contrast, the U.S. Copyright Office said that AI-generated images are “not the product of human authorship.” Platforms like Amazon already require self-published authors to disclose the use of AI in their writing, which the U.K.’s Society of Authors said will “ultimately benefit human authors and their readers.”
