一位作者表明立场，对阅读由大语言模型生成的虚构文学作品持保留态度。这位作者主张，对于2020年代及以后的任何虚构作品，"我想知道这篇文本是由一个人类以自己的语言创作的"[1]。
作者的反对理由基于对创意写作本质的理解。阅读他人作品能够改变读者自身的写作风格和统计特征[1]，这一过程对于文学创作尤为关键。与此相比，LLM的写作遵循统计中位数分布，倾向于"正常"的表达方式[1]，这与追求创新和个性化表现的文学阅读目标相冲突。作者指出，文学创作比技术写作或非虚构作品更强调创意和情感表达[1]，因此LLM生成的内容难以满足文学阅读者的需求。
An author has expressed resistance to reading literary works generated by large language models, arguing that fiction created by humans offers distinct value that AI-generated alternatives cannot replicate.[1] The author contends that "for any piece of fiction that I read from the 2020s and beyond, I want to know 'this text was written by a human being in their own words.'"[1]
The core objection centers on how LLMs approach composition. According to the author, language models generate text by following statistical medians, which leads toward conventional and unremarkable expression rather than innovative storytelling.[1] This statistical tendency, the author suggests, directly conflicts with what literature aims to achieve—the cultivation of unique voices and creative perspective.[1] By contrast, reading works authored by humans exposes readers to distinctive writing styles and statistical patterns that can reshape and elevate their own writing capabilities.[1]
The author further distinguishes between different types of written content, noting that creative fiction demands innovation and emotional authenticity in ways that technical writing or nonfiction do not.[1] This distinction underscores why the authorship question—whether a work originates from human effort or algorithmic output—matters significantly within the literary context.