Concept Craft | Part 1: Preface and Introduction

A Poetics of AI Art

Michael Filimowicz, PhD

Published in

Higher Neurons

8 min readFeb 2, 2024

This article series is the open access version of my eBook Concept Craft: a Poetics of AI Art.

Preface

There are many kinds of art that can be created using AI technologies, in fact, every kind of art! And not only complete art forms, but one may use AI technologies to produce elements for art, such as generating virtual 3D shapes that can be used in video games, sculpture, product design or architecture. AI art can be spatial or temporal, static or dynamic, interactive or generative and relational to any kind of spatiotemporal and material expression.

In this book, I will keep the discussion centered on prompt writing for making 2D images using platforms such as Stable Diffusion, DALL·E 2 and Midjourney. By narrowing the scope in this way and establishing a specific paradigm, the book can delve deeply into the intricacies of this particular form of AI-generated art, allowing for a more focused and comprehensive exploration.

By restricting the focus to this specific form of AI art making, which at the time of writing is the most popular way AI tools are used to make art, I aim to contribute to establishing a paradigm within the field. By thoroughly examining the poetics of prompt-based AI 2D image generation, I can outline the theoretical foundations, artistic considerations, and creative potential inherent in this particular method.

This focused exploration can then set the stage for future discussions, debates, and advancements in the field of AI-generated art, providing a valuable framework for further exploration, theorization and experimentation.

Introduction

A Performative Theory

Performative theory, also known as the performative turn or performative paradigm, is an approach within the humanities and social sciences that emphasizes the performative nature of language, discourse, and social practices. It goes beyond merely describing or analyzing phenomena, instead recognizing that theory itself can actively shape and produce reality.

At its core, performative theory challenges the traditional view of language as a neutral medium for conveying information about the world. Instead, it posits that language is performative — it doesn’t simply reflect reality but actively participates in constructing it. The performative power of language lies in its ability to not only represent but also create social realities, identities, and power dynamics.

The origins of performative theory can be traced back to several influential thinkers. One of the key figures in this regard is the philosopher J.L. Austin. In his How to Do Things with Words (1962), Austin introduced the concept of speech acts. He argued that utterances not only convey meaning but also have the power to perform actions and bring about changes in the world. For example, when a judge says, “I pronounce you guilty,” the utterance itself constitutes an action and has real consequences.

Building upon Austin’s ideas, the Judith Butler expanded the notion of performativity in her groundbreaking work on gender and identity. In Gender Trouble (1990), Butler argued that gender is not a fixed biological essence but a performative effect of reiterated social actions and speech acts. According to Butler, gender is produced through the repeated performance of gender norms and expectations, rather than being inherent or natural.

In addition to Austin and Butler, other theorists have contributed to the development of performative theory. Michel Foucault’s concept of “discursive practices” emphasizes the role of language in shaping knowledge, power, and subjectivity. Pierre Bourdieu’s theory of social fields and habitus also incorporates performative elements by highlighting how language and symbolic practices contribute to the reproduction of social structures.

Using generative AI tools to co-author a theory of the poetics of AI art can be seen as performative in the context of performative theory. Performative theory posits that language and discourse not only describe or represent reality but actively shape and produce it. In this case, by employing generative AI tools to co-author a theory, I am actively engaging in a performative act that enacts the subject matter I am theorizing about.

Active construction of reality: The act of co-authoring a theory of the poetics of AI art using generative AI tools demonstrates the performative nature of language and theory. Instead of passively observing or analyzing AI art, I am actively participating in its creation and shaping the discourse around it. The generative AI tools become a medium through which I perform and engage with the subject matter.

Blurring authorship boundaries: The use of generative AI tools challenges traditional notions of authorship and expertise. By collaborating with AI, I am disrupting the idea of a singular authoritative author and instead embracing a collaborative, co-authored approach. This challenges established power dynamics and opens up new possibilities for knowledge production and creative expression.

Performing the poetics of AI art: Generative AI tools themselves embody the poetics of AI art. They create, manipulate, and transform data to generate new artistic expressions. By employing these tools in the process of co-authoring a theory, I am not only describing or analyzing the poetics of AI art but actively participating in its practice. The act of using AI tools to generate content mirrors the very essence of the subject matter I am theorizing about.

Reflexivity and self-reflection: The use of generative AI tools for co-authoring theory creates a reflexive loop between theory and practice. As I generate content and engage with the AI tools, I constantly reflect on the implications of my thoughts and actions, the creative process, and the theoretical framework I am developing. This reflexivity enhances the performative nature of the endeavor, as theory and practice are intertwined and continually influencing each other.

By using generative AI tools to co-author a theory of the poetics of AI art, I actively embody the principles of performative theory. The act itself becomes a performative act that goes beyond mere description or analysis, actively shaping and participating in the subject matter being theorized. It blurs authorship boundaries, performs the poetics of AI art, and fosters reflexivity, thus exemplifying the performative nature of language, theory, and artistic practice.

A Performance

Welcome to this inquiry, Concept Craft: a Poetics of AI Art. Within these pages, we embark on a journey, delving into the theoretical frameworks that help explicate the production of AI images through the technique of prompt writing. This book invites you to explore the intricate interplay between human creativity and the emerging landscape of artificial intelligence, unveiling a tapestry of ideas and concepts that redefine artistic expression.

Concept Craft merges multiple disciplines, weaving together art, philosophy, cognitive science and cutting-edge technology. At its core lies the concept of a poetics of prompt writing as a gateway to the production of an infinite variety of aesthetics. Through this lens, we witness the convergence of artistic intention and the generative capabilities of AI systems, fostering a symbiotic relationship that unleashes an unparalleled range of visual possibilities.

Beyond the initial discussion of the distinctions between poetics and aesthetics, we encounter the Hegelian concept of material thinking. The transformative power of ideas takes center stage, as artists are propelled to explore the dialectical interplay between intention and materiality. Drawing inspiration from Hegel’s philosophy, we see how concepts become alive, shaping the AI-generated images that emerge from creative exploration.

Heidegger’s concept of poesis encompasses the creative and transformative nature of artistic production, while his ideas regarding thinking as a handicraft underscore the deliberate and skillful process of creative thought. Poesis, according to Heidegger, involves the act of bringing forth something into existence, emphasizing the artist’s capacity to evoke novel possibilities and shape the world through their artistic endeavors.

Thinking as a handicraft, on the other hand, accentuates the intentional and artisanal aspects of creative thought, drawing an analogy between the craftsmanship of a skilled artisan and the thoughtful construction of ideas. The interrelation between these concepts lies in the recognition that artistic creation necessitates both the imaginative and transformative power of poesis and the purposeful and dexterous thinking akin to that of a craftsman.

By conceiving thinking as a handicraft, Heidegger underscores the active involvement of the artist in shaping their artistic vision and the meticulous craftsmanship of ideas that contribute to the act of poesis, thereby amalgamating the creative and contemplative facets of artistic production.

Intertwined with these philosophical underpinnings of Hegel and Heiedegger, Concept Craft assimilates insights from Deleuze and Guattari, illuminating the nature of concepts themselves. We delve into the notion that an AI prompt becomes a form of Deleuzean assemblage, an intricate web of ideas, associations, and cultural references.

This allows us to frame the emergence of novel connections and unforeseen possibilities, as the AI system reworks the artist’s intentions, transcending traditional notions of creative expression whereby artists manipulate tools to realize their visions, since intelligent tools are not manipulable in the same way as camera buttons or paint brushes!

Eleanor Rosch’s ideas on fuzzy logic and family resemblance refine our practice of Concept Craft. We examine the nuances of category formation, where the boundaries between concepts become blurred, allowing for the emergence of a rich network of interconnectedness. Through fuzzy logic and family resemblance, the AI-generated images embody shared traits while retaining their unique aesthetic qualities, leading us to question and reimagine the very nature of categorization and artistic representation.

Within the process of creating AI-generated images, hermeneutic slippage emerges as a profound phenomenon, challenging our preconceived notions of meaning and interpretation. The dynamic interplay between prompt writing and image output brings forth unexpected divergences and multiplicity of meanings, inviting us to engage in a dance of interpretation. The veils of certainty are lifted, inviting us to embrace the fluidity of creative expression and the power of individual perception.

Finally, Concept Craft presents the groundbreaking notion that AI-generated art through prompt writing is a new form of neo-conceptual art. It transcends the traditional boundaries of creation, questioning established notions of authorship, interpretation, and the nature of concepts themselves. Within this framework, we witness the emergence of a transformative aesthetic experience, where the fusion of human imagination and AI capabilities creates a tapestry of possibilities that stretches the limits of our imagination.