How AI impacts subjectivity of literary creation

By YANG NING / 07-11-2024 / Chinese Social Sciences Today

The involvement of AI necessitates a reevaluation of the importance of the public subject in literary creation, challenging the traditional individual-centered view of creation and expanding our understanding of the possibilities of literature. Photo: TUCHONG


At the end of 2022, ChatGPT, an artificial intelligence developed by OpenAI, captured global attention. As a large language model, ChatGPT utilizes transformer architecture, a deep learning technology, to exceed expectations in various natural language processing tasks. ChatGPT not only generates natural and fluent conversations, but also drafts literary works such as poems, essays, and novels upon user request. Considering the powerful functions demonstrated by ChatGPT and its continuous iterations and development, it is necessary to reflect on the subject of literary creation.


Crisis of subjectivity: 

Another ‘end of literature’

The impact of AI writing on literary creation inevitably reminds us of the “end of literature” theory proposed by the famous American postmodern theorist J. Hillis Miller over 20 years ago. Miller believed that scientific and technological advances have changed people’s intellectual lifestyles and the environment in which literature thrives, heralding the decline of traditional literary studies. Looking back at the “end of literature” theory from over 20 years ago, it is clear that literature has not been replaced by another medium as Miller predicted. On the contrary, the emergence of network media has given rise to a new form of literature—online literature—which has remained popular to this day. As it turns out, literary works have become integrated into daily life in a more popular and accessible way.


In recent years, AI has produced an effect akin to a new “end of literature.” If AI texts become indistinguishable from human creations, the subjectivity of literary creation will face a profound test. Once the ability to create is not the feature which separates humans from other species, then what is the importance of literature and all art forms? Building on Miller’s theory, the new “end of literature” theory is novel for the following reasons. Miller asked whether linguistic text would be replaced by visual images, this is essentially a media transformation question in the digital age. In contrast, the new “end of literature” prompted by AI focuses on whether human literary creation will be replaced by AI. In this example, the creative subject transforms. The former explores whether the linguistic medium will end, while the latter investigates whether the creative subject will end.


Two attitudes: 

Subjectivity and non-subjectivity

At present, the academic community in China holds two main attitudes towards the crisis in literary creation brought about by AI writing. One group argues that no matter how advanced AI becomes, it can never replace the dominant role of humans in literary creation. The other group suggests that the development of AI might comprehensively challenge, or even replace, human creative subjectivity. In the end, these two attitudes reflect two different views on subjectivity and non-subjectivity.


According to traditional literary theory, literary creation is seen as an exclusive human ability—primarily because it requires creators to possess special psychological capabilities. This view of literary creation presupposes a fundamental premise: “anthropocentrism.” Under this premise, the uniqueness of literary creation is considered a manifestation of essential human capabilities. Therefore, in the face of the potential crisis of subjectivity triggered by AI writing, many scholars remain optimistic, believing that although AI brings challenges to literary creation, the structure of machines and algorithms means that their creations will never surpass human literary works.


In opposition to this optimistic attitude are other scholars who adhere to an “anti-anthropocentric” stance. These scholars believe that AI already threatens and might even fully replace human creative subjectivity. This concern stems from two facts. First, AI’s development trajectory is unknown, and even its developers are not fully aware of its intrinsic creative logic. Second, AI writes by first receiving input information and then producing output, this process is similar to how humans acquire knowledge and then create. This similarity highlights AI’s advanced capabilities in mimicking human thinking and creative processes. These two points are rooted in profound intellectual traditions, specifically reflected in the confrontation between the philosophies of subjectivity and non-subjectivity.


Where lies the subject: 

Between intention and text

The aforementioned schools of thought lead us to reflect on the following questions: What exactly constitutes the subject of literary creation? At what state can AI writing possibly possess subjectivity?


In traditional literary creation theory, imaginative thinking plays a crucial role. Thanks to imaginative thinking, inspiration, intuition, emotion, and unconsciousness interact and collectively drive the creative activity. However, until recently, academic investigations into literary creative thinking have mostly focused on the distinction between imaginative thinking and logical thinking, analyzing the uniqueness of literary creative thinking within the binary framework of rationality and sensibility, image, and logic. Scholars have not yet defined the essence of imaginative thinking on an ontological level.


In traditional literary theory, the process of literary creation is the transformation of human emotions and intentions into text. Following this logic, only AI which possesses emotions and intentions could potentially engage in literary creation. The next question is, can AI writing generate emotions and intentions? Although the question of whether AI has “self-awareness” has not reached a definitive conclusion, it has sparked another debate about AI writing. This debate reflects on a key question about the subject of creation: Is it important for the creative subject to have emotions and intentions? When facing literary works, readers can only experience, feel, and even speculate on the author’s creative intentions and emotions through the text, while the actual emotions and intentions cannot be directly perceived. Similarly, the focus of AI writing is on the text, and text analysis should become the starting point for discussing AI writing. The key to determining whether AI writing has emotions and intentions is not whether it truly possesses human emotions but whether it can simulate texts that have emotions and specific intentions, and whether these characteristics directly relate to the aesthetics of literary works.


Therefore, the key to studying the relationship between the subject of literary creation and the subject of AI writing lies in revealing the isomorphism between AI and literary creation. A deep exploration of this issue must consider two core questions: First, what degree of isomorphism exists between thinking and language? Second, where do the aesthetics of literary works specifically originate from?


Creative subject: 

Between thinking and language

First, let’s review the relationship between thinking and language. In AI writing, machine learning technology can be used to generate and optimize text data. The principles and processes of AI writing have distinct empirical characteristics. Its writing is primarily based on the analysis and summarization of phenomena, with less emphasis on abstract generalization or elevation. AI learns the intrinsic rules of language expression by analyzing vast amounts of text data, a similar process to the one human writers use to learn and master literary expression. Whether the writers are AI or human, their creation must adhere to certain language rules, such as grammar, sentence structure, and word choice.


Therefore, AI writing and human creation both have “heterogeneous isomorphism” relationships. When considering the text alone, there is no fundamental difference between the texts generated by AI writing and those created by human writers. When readers cannot distinguish whether the text before them is created by a human or AI, the emotions and intentions contained within the text possess a certain degree of subjectivity. In some cases, AI can even generate segments of text that are difficult for humans to understand. These segments can, to some extent, be seen as a form of “defamiliarization” in expressive practice, further expanding the rich possibilities of language expression.


Aesthetic subject: 

Emphasis on publicness

Next, let’s review the aesthetics of literary works. Where do the aesthetics in literary works come from? Traditional literary theory analyzes two elements: the creative subject and the text’s format. From this perspective, the aesthetic quality of literary works is believed to stem partly from the author’s emotions and intentions and partly from the special use of language. However, the emergence of AI writing indicates that the creative process is not simply a matter of transforming emotions and intentions into text. Solely relying on emotions and intentions doesn’t guarantee the aesthetic value of literary works. Similarly, focusing solely on the textual form doesn’t produce aesthetic appreciation. The aesthetics of literary works arise from multiple factors, and in this formation process, the publicness of aesthetics is more important than the privateness.


“Aesthetic publicness” refers to the universality and resonance presented by the aesthetic value of literary works. It transcends the intentions and subjective emotions of individual creators and involves broader social, cultural, and shared human emotional experiences. This publicness manifests in two ways: First, the work will resonate among different readers, and second, the text will interact with social, cultural, and historical contexts.


Analysis of the aesthetic quality of literary works should shift from a traditional approach that narrowly considers the author’s emotional experience, language expression, and personal talent to a broader focus on the commonalities of history and culture. In this way, literature is seen not only as the product of individual creation but also as part of a broader social and cultural dialogue, highlighting its deeper publicness and universality.


In essence, the relationship between AI and human writers can be seen as the relationship between public and private subjects. AI, as a public writing subject, bases its writing process and output not on unique individual experiences, but on the comprehensive processing of vast amounts of data and algorithms. This data and algorithm-based creation method emphasizes the value of public knowledge and public experience in literary creation. The involvement of AI necessitates a reevaluation of the importance of the public subject in literary creation, challenging the traditional individual-centered view of creation and expanding our understanding of the possibilities of literature.


Yang Ning is from the School of Liberal Arts at Minzu University of China.


Edited by YANG XUE