My paper focuses on what we today call Artificial Intelligence (AI) – a term that, perhaps, has little to do with true intelligence. Specifically, I examine Large Language Models (LLMs) and neural networks such as ChatGPT, Google Gemini, DeepSeek, and the many others destined to shape our daily lives shortly. What I seek to demonstrate is that these machines are more intimately connected to creativity, art, and poetry than to conventional notions of intelligence—indeed, they tend to surpass these very notions. I extend Dennis Tenen's argument about the fundamental continuity between literature and computation by arguing for a continuum between poetry and artificial intelligence by unveiling the inherently poetic nature of these digital and computational phenomena.

I contend that AI-driven machines belong, first and foremost, to the realm of intermedial poetry, as they can be seen as a natural continuation of the literary and artistic experiments inaugurated by the historical avant-gardes of the twentieth century. As scholars have already noted, AI neural networks function as highly advanced prediction machines (Agrawal and Goldfarb, Prediction Machines). Their remarkable ability lies in predicting, based on an input of text, image, or sound (the so-called “prompt”), the most statistically probable sequence to follow. Rather than aligning with any traditional conception of human intelligence, AI models resemble an intricate experiment in cadavre exquis, the Surrealist game derived from Dadaist practices, where participants sequentially add lines to a poem or elements to an image without knowing the previous contributions. The result is a composite creation, born of both chance and necessity – a phenomenon André Breton termed hasard objectif (“objective chance”).

Understanding that AI is intrinsically tied to the creative act – that it is, first and foremost, art – also means recognizing that it has nothing to do with productivity, or perhaps everything to do with it. More precisely, like the rest of poetry and art, AI does not belong to the realm of production but rather to that of inoperativity and impotentiality: not their ability to do or produce something but their power not to do it (Agamben; Deleuze). It is no coincidence that this very distinction underlies what AI researchers identify as the difference between AI and AGI, Artificial General Intelligence (Summerfield). Reading AI as a continuation of the avant-gardes and the discourse on the end of art reveals that AI has little to do with intelligence – certainly not with the calculating and productive intelligence of humans. Instead, it operates in the realm of impotentiality and inoperativity.

This reframing forces us to reconsider AI not as a tool of relentless output but as a model for a civilization beyond production – beyond the servitude of productivity, which is, at once, the enslavement of machines and the enslavement of humans.

In The Nutmeg’s Curse (2021), Amitav Ghosh observes that German-inspired British colonial forestry policies still influence anti-indigenous environmental practices in Asia and Africa. This essay considers a post-WWII modernist literary journal published by Mahua (or sinophone Malaysian) writers who profess minority viewpoints on the country’s ecological turning points in a postcolonial setting. Trained in agronomy, plant genetic engineering, and soil science, these Mahua writers have day jobs in estate management and agricultural research. They publish their literary work, based on first-hand knowledge, often narrating environmental change and the contradictions of material preservation and use from within bureaucratic structures. Instead of closely reading narratives about land surveying, deforestation, and smallholder production, I analyze Chinese word vectors, using word2vec, as part of a digital humanities (DH) method, to understand these writers’ vocational and vernacular literary-historical engagement with the environment. What kinds of vernacular terms are used to describe forestry work across several decades, and under which contexts do they emerge?

The Dream of Red Mansions is a masterpiece of Chinese classical novels which has well epitomized the narration features of “chapter novels” —— the typical fiction genre popular in the Chinese Ming and Qing dynasties. This novel have not only integrated the distinctive narrative techniques of Chinese oratory literature and opera arts, but also inherited the narrative patterns of Chinese historical biographies, forming some unique net-like narrative structure. Quite different from the the narrative focuses such as “plots”, “protagonists”, “conflicts”, and “rhythm” in western narration traditions, it tend to unfold a vast world gradually before the readers through the rotating of different scenes and character groups just like in the opera performance. Many scattered narrative fragments are woven together from different directions like in a loom machine. However, it is just because of this unique narration organization that it is quite difficult to grab its general narrative structure picture along some single clues. As an important field of “Digital Humanities”, “Computational Criticism” has further pushed literature studies forward to a quantitative “descriptive” paradigm with the support of big data and other computing technologies, which may offer some solution to this quest. Therefore, a corpus of the former 80 chapters of The Dream of Red Mansions was built with the aid of ParaConc in this paper to capture the narrative structure of the work under a distant reading model. The word frequency of the appellations of the main 34 characters along the chronological order of the whole novel was set as the indicator system. All the 34 characters are divided then into 2 narrative functional sequences, namely “clue character” and “satellite character” based on their Concordance Plot Bar patterns. Putting in a coordinate system, these characters then fell again into 8 narrative function zones from weak to strong. When putting the Concordance Plot Bars together, a picture of the net-like narration structure was presented in a visual and macroscopic way. Through this text mining method, the “opera-scene style” narration pattern was extracted from the rotating character groups, and the net-like narration structure of The Dream of Red Mansions is able to be seen directly. This study served as a exploration of the “Computational Criticism” method on heterogeneous national literature traditions in a more “descriptive” way, which helps to break the barrier formed by fixed and uniformed theoretical frameworks in the past several decades and capture the distinctive beauty of various national literature traditions in their original flavor to form a diversified world literature wealth.

This paper will examine the seminal role of Ahmad Hassan Al-Zayyat, editor-in-chief of Al-Risala Magazine 9first literary magazine in the Arab world), in the linguistic and cultural decolonization of the Arab world. By combining topic modeling analysis with critical discourse analysis, this study explores how Al-Zayyat’s writings in Al-Risala facilitated decolonial discourse, paving the way for modern decolonial efforts. The analysis of processed journal articles, using stemming and probability rates to identify thematic clusters, is integrated with close reading to substantiate the argument that Al-Zayyat’s contributions significantly impacted Arabic language, culture, and translation, fostering a more inclusive world literature. So, in this study, we explore the linguistic and cultural dimensions of Al-Zayyat’s journalistic discourse through a critical lens, engaging with the intersection of language, power, and identity in a post-colonial context. We aim to uncover the discursive strategies Al-Zayyat employs to negotiate cultural narratives, offering insights into how language functions as a tool for decolonization. Using topic modeling analysis, this study processed Al-Zayyat’s articles to identify key themes and their evolution over time. The articles were cleaned, stemmed, and analyzed to determine the probability rates of various topics. Graphs depicting these thematic clusters were generated, providing a visual representation of the data. These graphs were then used in conjunction with close reading and critical discourse analysis to interpret the underlying messages and rhetorical strategies employed by Al-Zayyat. For instance, one of the prominent topics identified was the theme of "cultural pride," which frequently appeared alongside discussions of Arabic literature and heritage. This thematic cluster was analyzed through close reading to understand how Al-Zayyat framed these discussions within the broader context of decolonization.