Author Haruki Murakami said his novels are distinct from AI-generated literature due to the spontaneous emergence of characters and narrative elements.
As generative artificial intelligence increasingly permeates the creative arts, the perspective of one of the world's most influential novelists provides a benchmark for the perceived gap between human intuition and machine learning.
Speaking in Tokyo, Murakami said his creative process is "completely different" from what AI can create [2, 3]. He said that in his writing, characters suddenly show up, noting that this is not something that comes out from analogy [2]. This spontaneity, he suggests, is a uniquely human trait that defies the predictive nature of current technology.
"AI probably can't do that," Murakami said [1].
His comments come as a new book hit shelves in Japan on Friday [3]. The release coincides with a global debate over whether large language models can replicate the "soul" or unpredictable spark of high literature. While AI can analyze vast datasets to mimic styles, Murakami said that the genuine birth of a character is an organic event rather than a calculated output.
By framing the difference as a matter of analogy versus spontaneity, the author suggests that AI is limited to rearranging existing patterns. In contrast, human creativity allows for the sudden appearance of elements that do not follow a predetermined logical path, a hallmark of Murakami's own surrealist style.
Throughout his career, Murakami has explored the boundaries between reality and the subconscious. His current stance positions the act of writing not as a data-processing task, but as an intuitive discovery of characters who possess their own agency within the narrative.
“"AI probably can't do that."”
Murakami's assertion highlights a fundamental tension in the AI era: the distinction between synthesis and creation. While AI can produce text that resembles a specific author's style through pattern recognition, Murakami argues that the 'spontaneous' emergence of narrative elements is a human cognitive process that cannot be simulated by algorithmic analogy.
