Research associate Li Guocheng from the School of Liberal Arts of Nanjing University has published "Artificial Intelligence Literature and Its Challenges to Modern Literature Theories" on Social Sciences in China. The paper examines the history, challenge, and potential of artificial intelligence literature from the perspective of post-literature and post-humanism.
Li's research first identifies artificial intelligence literature as a new form arising from the intersection of literature and technology. It is categorized into two types and stages, respectively sequential paradigm and connected paradigm based on different technological approaches. Recent breakthroughs in this field are attributed to rapid advancements of deep learning and neural networks brought by the AI technological conversion. However, historically, the mechanism of AI literature has already sprouted in former human literary activities and resonates with various avant-garde literary experiments of the 20th century, in that the fundamental features of AI literature like automaticity, randomness, and procedure echo with the purpose chased by "automatic writing" of surrealism, "cutting technique" of Dadaism, and "constraining writing" of OuLiPo. To some extent, AI literature responds to the inner appeal of these avant-garde literary experiments. In the face of the crisis of traditional literature in the beginning of the 20th century, various sorts of literary effects designed by these avant-garde literature movements can be achieved more conveniently and richly with the condition of computer and AI technology. This suggests that AI literature should not be depreciated outright in terms of its significance and potential.
The paper then addresses various critics against AI literature based on modern literary theories. AI literature is a new phenomenon emerging in the late 20th century under the circumstances of post-literature and post-humanism, but evaluations often rely on standards of author-centred perspective established in the 17th and 18th centuries, indicating a notable misalignment. In fact, modern literature theories are just a special understanding of literature under certain historical conditions instead of revealing the invariable essence of literature. In the middle and late period of the 20th century when AI literature came into being and started to develop, modern literature theories encountered severe clashes, leading to endless statements like "the death of the author", "the end of literature". Li's article discusses three core points of contention regarding AI literature under modern literary theories: emotionality, inspiration and authorship.
Emotionality, which is thought to be difficult for AI to simulate, is often viewed as a fundamental element of modern literarture. However, the influential literary objectivism of the 20th century has advocated weakening or even excluding the role of emotions in literary creation and research, and Alain Tomé’s research has also pointed out that the "theory of emotional expression" in modern literary concepts contains a kind of fallacious reasoning, from the emotional characteristics of the text to the author’s emotional experience arbitrarily attributed. If emotions should only be expressed and grasped in the textual characteristics of literature, then it provides space for AI to deeply study and imitate the emotional characteristics of literary texts.
Inspiration, which, together with creativity, is highly valued by modern literature values, but not all works have emerged solely from genius and inspiration when inspecting the history of human literature. Early literary works, such as Homer’s Epics, were collaborative narratives rather than products of individual inspiration. In the 20th century, the concept of “weaving” has been resurrected, and writers have once again been promoted and positioned as weavers of various writing.
Authorship. The issues of emotionality and inspiration lead to author-related questions, which is essential to the controversy over AI literature. The ways in which authors create have historically been influenced by technological conditions. As a result, modern literary theories shaped by typography do not align with the characteristics of verbal communication and digital revolution.
Moreover, the paper argues for a re-examination of AI literature from a post-human perspective. The reason why modern literary concepts are seriously challenged by AI literature is that the author-centralism on which it is based is a kind of anthropocentrism, while AI literature is the product of the post-human condition. This shift requires moving from anthropocentric viewpoints to a more realistic cyborg perspective that embraces human-machine collaboration. It is generally believed that the "Turing-style artificial intelligence literature" is fundamentally a deformation of anthropocentric discourse, and it has not yet been realized under the current technical conditions; what the current technology has achieved is still "Liklade-style artificial intelligence", which is subordinate to the cyborg discourse. It is a decentralized cooperation and co-creation between man and machine. If we look at the entire process of literary creation activities, paying attention to not only the autonomy of AI, but also face up to its dependence on people, then AI literature can be said to be a kind of decentralized "cyborg literature." It is not the self-expression of the subject of AI, but a brand new form that matches the symbiotic cyborg information system of man and machine. From this perspective, AI literature is no longer a botched imitation and potential threat generated by another literary subject, but as an expression of complete cyborg human abilities, it can significantly expand the conceptual space and forms of literature, offering cross-boundary potential between human and machine, authors and readers, as well as linguistic and non-linguistic texts.
Meanwhile, the paper emphasizes that while the "post-literature" characteristics of AI literature render it unsuitable for evaluation through traditional literary theories, it is still under the instruction of normative literary frameworks. A "post-human" approach towards AI literature seeks to engage with human existence in a broader context, striving for a better future for humanity within the evolving historical landscape. As the product of the "language machine" criticized by Heidegger, AI literature becomes the "double-sided god of Janus" that combines danger and opportunity, and requires us to constantly question and reflect on its development.
Writer: Bai Jingwen
Reviser: Ma Jinge
