Artistic Strategies of Mario Klingemann as a Cooperation Between Humans and Artificial Intelligence Systems.

Annotation

The paper analyzes the artistic strategies of Mario Klingemann between 2018 and 2024, drawing on a qualitative analysis of interviews with the artist and his theoretical reflections on AIArt. Klingemann’s work reflects the posthuman era, redefining the relationship between humans, technology, and authorship in the artistic discourse. Situated within an interdisciplinary framework, it integrates elements of database aesthetics, algorithmic culture, and the concept of artificial intelligence as a co-author. The research identifies key methodological approaches, particularly the demystification of algorithmic mechanisms and transcoding, which enables translation between cultural and technical formats. Additionally, it explores distributed agency, where the creative process emerges as a synergy between human and non-human agents, with AI acting as a creative partner. In the context of the posthuman era, Klingemann redefines authenticity and originality as a dynamic outcome of collaboration between humans and AI, thereby opening new epistemological and aesthetic perspectives. The study provides a theoretical framework for examining agency and cognitive processes in AI systems, reflecting fundamental shifts in the perception of artistic creation and creativity in the digital age.

Introduction

This paper analyzes the artistic strategies of Mario Klingemann between 2018 and 2024, drawing on a qualitative analysis of interviews with the artist, his theoretical reflections on AIArt, and artistic strategies utilizing artificial intelligence. Klingemann’s work represents a critical reflection on the posthuman era, redefining the relationship between humans, technology, and authorship within contemporary artistic discourse. Simultaneously, his approach aligns with the broader theoretical perspective of the AIArtworld concept, which reflects new forms of collaboration between artists and artificial intelligence systems.

The definition of AIArt operates within a complex interdisciplinary framework, where scientific, technological, philosophical, and ethical factors play a crucial role. This dynamic character of AIArt integrates elements of database aesthetics, algorithmic culture, and the concept of artificial intelligence as a co-author. Despite the rapid development of generative AI, the definition of AIArt remains unstable and subject to discussion, influenced by issues of originality, copyright, and the perception of artistic value in the digital era. The qualitative research in this study focuses on identifying Klingemann’s artistic-research strategies, with demystification emerging as a central methodological approach — unveiling the algorithmic mechanisms that shape the visual aesthetics of generative models. This demystification is closely linked to transcoding, translating cultural and technical formats into new artistic forms, synthesizing diverse media and aesthetic codes. Klingemann’s strategies utilize aesthetic and stylistic elements to explore the role of humans in a technologically interconnected world.

Additionally, his artistic projects reflect the concept of distributed agency, where the creative process is not solely attributed to human agency but emerges from the synergy between human and non-human actors. This concept is rooted in theoretical discussions in philosophy of technology, posthumanism, artificial intelligence, and social sciences. Distributed agency challenges traditional anthropocentric notions of agency, presenting it instead as a decentralized and distributed process, where agency is shared among various technological and biological entities.

These artistic strategies fundamentally redefine the relationship between the artist, the medium, and technological systems, serving not only as a methodology for artistic creation but also as critical commentaries on the interaction between technology and human creativity. Approaches based on the principles of transcoding emphasize the translation between cultural and technical formats, enabling new models of collaboration between humans and artificial intelligence. This concept holds significance not only in artistic practice but also in broader societal discussions on the impact of algorithmic systems on creative processes. In the context of the posthuman era, Klingemann redefines authorship and authenticity as a dynamic outcome of collaboration between human and AI actors, opening new epistemological and aesthetic perspectives. His work provides methodological frameworks for exploring agency and cognitive processes in AI systems, expanding interdisciplinary discussions on artificial creativity.

Klingemann’s artistic projects reflect these ideas through the concept of interactive AIArt, where AI does not function as a passive tool but as an active creative partner. Within this human-algorithm interaction, authorship is dynamically redistributed, and the artwork ceases to be the product of a singular artistic intention, becoming instead the result of iterative processes between the artist, algorithms, and the audience. This model leads to a fundamental redefinition of originality, authenticity, and the perception of artistic works in the digital era, highlighting a posthuman understanding of creativity as an emergent phenomenon that transcends the traditional boundaries between biological and technological entities.

AIArt and Artistic-Research Strategies

An Attempt to Define AIArt

AI Art — Artificial Intelligence Art is part of the broader evolution of digital art, in which generative algorithms, particularly through machine learning, become key tools for artistic creation. This form of artistic production can be defined on multiple levels. First, AI Art can be understood as a continuation of the development of software tools that enable new ways of generating visual content, music, and text, analogous to the historical influence of digital creative tools. Second, AI Art functions as the outcome of database-driven processes, where artificial intelligence operates through selection and combination of elements from vast image databases, generating new hybrid forms in accordance with the concept of database aesthetics. Third, AI Art is embedded in algorithmic culture, where it is not perceived merely through its final artworks but primarily as a process determined by the structure of data sets, machine learning methods, and programmer decisions. A much more fundamental defining characteristic is the perception of artificial intelligence as a co-author. Rather than considering AI as an autonomous creator, this perspective emphasizes its role as a creative partner that extends the spectrum of human creativity, similar to other digital tools. This set of characteristics shapes a historically unprecedented relationship between humans and machines, one that possesses the materially realizable potential to operate with manifestations of consciousness. The prediction and research of the material realization of so-called artificial consciousness is increasingly becoming a subject of artistic strategies and experimental practices.

Lev Manovich and the theoretical frameworks of AIArt

Lev Manovich, in his study Defining AI Arts (2019), presents three approaches to defining AI art, each reflecting different epistemic and aesthetic aspects of this domain. The first approach is the concept of the “Turing Test for AI Art,” which assumes that if artificial intelligence systems can generate artifacts visually indistinguishable from human-made works, it raises questions about the convergence of historical and contemporary artistic outputs. This perspective suggests that AI Art can embody the “immortality” of past artistic styles, as AI models are capable of generating works in the visual language of old masters while lacking the authentic artistic context of their creation. This challenges traditional notions of authorship and originality, positioning AI-generated works within a framework that is both aesthetic replication and computational innovation.[1]

The second approach focuses on the distinction between traditional computer art and machine learning methods, with Manovich highlighting the conservative nature of many generative models. According to him, AI Art primarily reproduces existing aesthetic patterns rather than creating radically new forms. This aspect is evident in models such as DALL·E or Stable Diffusion, which draw from large datasets of historical artworks, effectively imitating past aesthetics instead of generating original structures.

The third approach considers AI Art as a potential tool for transcending traditional human aesthetic patterns. Manovich formulates the thesis that the future of AI Art does not lie solely in simulating existing styles, but rather in the ability of artificial intelligence to generate new visual and auditory structures that are both appealing to humans and radically unpredictable. He emphasizes the need for a shift from imitation to innovation, where AI could redefine the boundaries of creativity by producing artworks that are not explicitly learned but emerge as a new aesthetic phenomenon.

Since 2021, Manovich, together with Emanuele Arielli, has been developing these ideas in a series of publications titled Artificial Aesthetics (2024), analyzing contemporary trends in AI Art. This series explores how generative models are redefining the creative process, addressing issues of authenticity, originality, and creativity. According to Manovich and Arielli, AI has not yet demonstrated the ability to autonomously redefine artistic movements, but rather functions as a tool for human creators, who use it to expand their aesthetic repertoire. The key question remains whether AI can become a truly creative subject, or whether it will remain merely a sophisticated mechanism for recombining existing aesthetic patterns.[2]

In May 2024, Manovich published a post on the professional network LinkedIn, presenting arguments in support of his claim that there is no specific artificial intelligence aesthetics. He bases this assertion on the premise that until the 20th century, art was fundamentally rooted in manual skills and technical training, and even average works by Old Masters contained visually appealing qualities, such as human faces, interiors, and landscapes. From this, he concludes that artistic value has historically been tied to technical competence rather than conceptual or abstract content. In his reflections, Manovich critiques contemporary artistic production, arguing that it operates without clearly defined criteria and lacks systematic training in technical skills. He asserts that contemporary art is often visually unengaging, as it does not incorporate the craftsmanship aspects that shaped the aesthetics of previous artistic epochs. This perspective extends to his discussion of generative AI, where he expresses skepticism about the ability of today’s artists to create exceptional works using artificial intelligence.

On the other hand, Manovich sees the potential of AI Art in professionals from the cultural industries, such as illustrators, designers, and photographers, who are already utilizing AI tools to produce visually innovative works. According to him, technical skills remain the key factor in enabling the production of aesthetically relevant outputs, rather than the mere existence of generative models. Through this argument, Manovich highlights a paradox in AI Art — on one hand, it allows the general public to experiment with artistic creation, yet at the same time, it does not guarantee aesthetic quality without a solid artistic and technical foundation.[3]

Manovich’s theoretical approaches to AI Art reveal fundamental epistemic and aesthetic tensions shaping the contemporary discourse in this field. On one hand, AI expands creative possibilities, yet on the other, it still functions primarily as a tool for reproducing existing styles. The key challenge remains overcoming this conservative tendency and reaching a stage where AI can generate radically new aesthetic paradigms. Through this, Manovich opens a broader philosophical discussion on the nature of creativity, authenticity, and the future of artistic practices in the age of artificial intelligence.

Artistic Research Strategies in the Context of Epistemic Shifts and Transpositions

The current academic discourse reflects the definitional uncertainty that characterizes the field of artistic research. This research is continuously shaped by interdisciplinary discussions, with its epistemic boundaries expanding due to the interplay between artistic and scientific methodologies. The key question remains whether artistic research generates new knowledge through its distinct artistic methods, or whether it primarily serves to open meta-research inquiries and explore alternative forms of knowledge representation.

The terminological differences within this field — such as artistic research, practice-based research, recherche-création, and künstlerische Forschung — suggest that the concept is conditioned by geopolitical and cultural contexts. Despite the frequent interchangeability of these terms, their varying definitions and interpretations fundamentally influence methodological approaches, development, and institutionalization of artistic research. Its formalization in the second half of the 20th century emerged as a response to the need for new epistemic strategies, enabling the integration of artistic and scientific processes within the broader academic discourse.

In his study Contemporary Research (2023), Michael Schwab identifies two fundamental epistemic shifts: the transition from traditional concepts of knowledge to more dynamic research paradigms and the shift from dominant forms of artistic production toward smaller, less formalized experimental practices.[4]

This development can be interpreted as a response to the increasing complexity of contemporary knowledge, where traditional scientific and artistic frameworks no longer provide sufficient analytical tools for a rapidly changing epistemic environment. Schwab, in his analysis, draws on the concepts of Hans-Jörg Rheinberger, particularly his ideas on experimental systems and graphematic space. He proposes that artistic research should not create fixed knowledge objects but rather produce sets of traces as proto-forms that remain aesthetically and epistemically indeterminate, opening new possibilities for transdisciplinary interactions..[5],[6],[7]

In this context, the significance of representation as the dominant form of knowledge is diminishing, a shift observable not only in the artistic field but also in the scientific domain. Artistic research is gradually transforming into a transdisciplinary platform, where different knowledge structures are synergistically interconnected. This shift challenges artists to create new expository forms of knowledge that transcend traditional representational frameworks, ensuring epistemic relevance amid the accelerating complexity of knowledge production. At the same time, interdisciplinary and transdisciplinary research is developing, often existing at the periphery of academic structures. This trend suggests a potential erosion not only of the instrumentalization of research but also of the very concepts of knowledge that constitute the foundation of contemporary scientific and artistic discourse. In this regard, Schwab speaks of an epistemic shift, wherein research does not merely generate new knowledge objects but redefines the very boundaries of what constitutes knowledge. He attributes this development to two fundamental global transformations — the acceleration of processes and the increasing complexity of knowledge.

This issue is closely linked to the discourse on multi-stable positions, which emphasizes the role of transpositions in artistic research. This concept can be interpreted through Duchamp’s neologism “infra-thin,” referring to the subtlest differences between identity and its differentiation. Thierry De Duve applies this concept to the interpretative ambiguity of Duchamp’s fountain, arguing that the “infra-thin” separation allows for distinguishing between two identities that appear identical yet are fundamentally different. This principle is also crucial in artistic research, where transpositional processes blur the boundaries between knowledge and its representation, generating new forms of epistemic articulation. Schwab acknowledges that artistic research operating through transpositions is not a stable field, discipline, or fixed concept, as each new example pushes its definitional boundaries further. However, this dynamic character is characteristic of an artistic practice that reflects the constantly changing epistemic conditions of contemporary knowledge.[8],[9]

Language as a tool for artistic research strategies

Generative models, such as Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs), play a key role in art creation based on existing data, forming the fundamental technological principles of AIArt. These algorithms utilize vast amounts of visual information to generate new image outputs, thereby mimicking the aesthetic qualities and compositional principles of both historical and contemporary art. One of the most significant factors influencing the output of generative models is language as an instruction.

Within the framework of prompt engineering — the process of formulating input text commands — language functions as a parametric regulator of the outputs produced by models like DALL·E or Stable Diffusion. The choice of words, the structure of the text input, and its semantic interpretation can fundamentally impact the final visual output. This relationship between linguistic input and generated imagery raises questions concerning interpretation, creativity, and the autonomy of artificial intelligence in the artistic process.In the context of generative art, there is also a tension between explicit and emergent meaning. A significant example is the work of Mario Klingemann, who deliberately experiments with manipulating linguistic instructions to achieve unexpected results. This approach demonstrates that while AI models operate based on probabilistic calculations, their outputs can transcend predictable patterns, revealing new aesthetic forms that emerge from the interaction between human input and generative systems.

Interactive AIArt represents artistic-research strategies in which AI acts as an active partner to the artist, responding to viewer inputs and generating outputs based on dynamic parameters. This approach to art is not solely based on verbal language but also incorporates visual symbols, with human-machine communication occurring through algorithmic interpretation. This model of artistic production exhibits parallels with performative forms such as happenings and conceptual art, where the artwork does not emerge as a static object but as a process in continuous interaction with the audience.[10]

In the context of AI art, language extends beyond mere words and code, taking on a broader philosophical dimension that reflects the relationship between meaning, interpretation, and authorship. Similarly, AI art can be understood as a space where the meaning of a generated image is not fixed but is shaped by the conditions of its creation and the way it is interpreted by the viewer.

This concept destabilizes the traditional understanding of an artwork as a definitive and authorially fixed product, instead opening it up to a realm of flexible semantics. This principle is evident in the strategies of Mario Klingemann, who, in projects such as Memories of Passersby I, experiments with manipulating AI systems to generate visual outputs that remain in a constant state of transformation and reinterpretation.[11]

Generative models produced by artificial intelligence do not operate with fixed meanings but instead create ambiguous and fluid visual structures that challenge the dichotomy between original and copy. AI-generated images are not the result of artistic intent in the traditional sense but rather the product of a decentralized process in which the model recontextualizes visual elements extracted from vast datasets. This principle finds parallels in Duchampian ready-made art, where the meaning of an object is not inherent but is generated through its placement in a new context.

A similar issue is reflected in the work of Anna Ridler, particularly in projects such as Mosaic Virus, where she employs AI to generate images of tulips, referencing both the historical tulip mania of the 17th century and contemporary questions of economic speculation on digital images. A semiotic analysis of visual signs offers another perspective on AI art, understanding it as a system of symbols akin to language. AI-generated art can be seen as an extension of visual language, where algorithms not only simulate existing styles but also create new forms of symbolism that may not always be immediately comprehensible to humans.

Thus, AIArt emerges as a hybrid linguistic phenomenon, combining interactivity, deconstruction, and the semiotics of visual signs. Its aesthetic value and meaning are in a constant state of formation, depending on the dynamic relationship between the artist, the algorithm, and the viewer. Contemporary AI art production reflects this shift from a static artistic object to an ever-evolving network of meanings, where the boundaries between authorship, interpretation, and technology remain open.[12],[13]

Mario Klingemann’s Artistic Research Strategies: Demystification, Transcoding and Redefinition of Authorship

Methodological Approach in the Artistic Research of Mario Klingemann

Mario Klingemann represents a unique case in the field of AIArt, as he redefines the creative process in which humans are not the sole agents, but rather, artworks emerge through collaboration between human and non-human entities. His artistic practice, developed outside traditional institutional art education, supports Lev Manovich’s argument that contemporary professional artists cannot be expected to produce exceptional outputs in collaboration with AI. Instead, Manovich sees real potential in experts from the cultural industry, who possess a specialized form of knowledge capital that can harmoniously coexist with artificial intelligence systems.

The aim of this qualitative study is to identify the artistic-research strategies of Mario Klingemann, with a particular focus on the principles of demystification and transcoding. These strategies redefine authorship and authenticity as a form of cooperation between humans and AI systems. At the same time, they highlight a specific approach to artistic research methodologies that engage with language as a tool of artistic inquiry.

This research employs qualitative methods of interpreting texts and visual materials. The primary sources include the interviews Creation, Curation, and Classification: Mario Klingemann and Emily L. Spratt in Conversation (2018) and Art Beyond Mechanical Reproduction: In Conversation with AI Artist Mario Klingemann (2024). The analysis is based on a thematic examination of key concepts within the context of artistic research, incorporating a critical reflection that attempts a discursive analysis of the redefinition of authorship and authenticity in AIArt. The study hypothesizes that Mario Klingemann’s artistic practices reflect a methodological transformation of art in the posthuman era. This research aims to provide a deeper understanding of the relationship between AI and artistic creation within the framework of contemporary epistemological shifts, contributing to a broader discussion on the future of artistic research and digital creativity.

Autodidact as an Advantage

Mario Klingemann’s journey into the realm of art and technology began in Munich, Germany, where he was inspired by a diverse range of influences from his surroundings at an early age. As an autodidact without formal education in art or technology, he was profoundly shaped by the technical fascination of his father, an engineer, and the artistic activities of his mother. Early exposure to computers and plotters in the 1980s sparked his interest in merging technological processes with visual creativity. Without access to advanced technologies, he experimented with home computers, resourcefully pushing the limits of the hardware available at the time. This early phase of experimentation laid the foundation for his later work, where art and technology became inseparable.

A pivotal phase in Klingemann’s career was his work in the advertising industry, which introduced him to modern graphic tools such as Photoshop and exposed him to the aesthetics of techno music and the visual language of the digital age. These experiences enabled him to develop a unique artistic approach that combines digital technology with innovative artistic practices. Klingemann’s work represents a significant contribution to the intersection of art and artificial intelligence, with his technical-artistic background serving as a distinctive foundation for his experiments with new media.

In May 2024, Lev Manovich published an argument on LinkedIn asserting that no specific AI aesthetics exist. He based his view on the historical premise that art until the 20th century was deeply rooted in skills and formal training. Manovich highlighted that even average works by old masters featured visually appealing elements, such as portraits, landscapes, or interior scenes, aligning with the aesthetic norms of their time. He contrasted this with contemporary visual arts graduates, whose work he argued often lacks clear technical criteria, resulting in art that fails to offer compelling visual content.[14]

Manovich further argued that current professional artists cannot be expected to produce exceptional AI-assisted works due to their insufficient technical skills. He saw greater potential in professionals from cultural industries, such as designers, illustrators, and photographers, who effectively use AI tools to create works comparable to traditional art by leveraging their technical expertise and understanding of aesthetics. According to Manovich, artistic talent, understood as a biological predisposition, might be partially replaced by strategic action and technological proficiency in the age of AI. This shift reflects a broader socio-educational transformation of the art world, reshaping the artist’s habitus. A prime example of this transformation is the work of Mario Klingemann, who uses technological tools to explore new artistic forms and concepts, illustrating the evolution of artistic identity in the era of artificial intelligence.[15],[16]

Demystification as an Essential Basis for Artistic Strategies

Mario Klingemann perceives art as the exploration of systems, structures, and the fundamental rules that govern them. He likens his creative process to the “demystification” of these rules, drawing parallels with the tale of The Emperor’s New Clothes. By unveiling the mechanisms behind cultural artifacts and artworks, Klingemann fosters a deeper understanding of their nature and implications. Central to his practice is the belief that art and technology co-evolve. This interplay not only expands the boundaries of artistic expression but also opens new avenues for intellectual and philosophical inquiry.

For Klingemann, the role of the artist transcends mere creation, encompassing curatorial activities that require selective processing and interpretation. His explorations raise fundamental questions about authorship and creativity in the digital age. By integrating machine learning into the artistic process, Klingemann challenges the notion of the artist as the sole creator. Instead, he envisions a collaborative dynamic between humans and machines, where the artist functions as a guide, curator, and interpreter of computational creativity.

Lev Manovich’s reflections on the socio-cultural influences shaping the habitus of the artist, while insightful, are not fundamentally novel. They align with Howard Becker’s definitional positions on the Artworld.[17] These ideas provide a framework for conceptualizing the emerging AIArtworld as a network of human and non-human actors, including organizations equipped with AI systems, whose activities are integral to the production of specific events and objects. Klingemann’s work exemplifies signs of this nascent AIArtworld, where the creative process redefines traditional artistic paradigms.

Projects like The Butcher’s Son (2017) and Circuit Training (2019) exemplify Klingemann’s strategies rooted in transcoding principles. These strategies explore the relationship between human-AI partnerships, where humans shape AI while simultaneously redefining their identity and role within a technologically interconnected AIArtworld. Transcoding — the translation of languages, including visual ones — facilitates a guided synergy between human creativity and machine capabilities. This process embodies a paradigm shift in authenticity and authorship, symbolizing the emergence of a new “A(I)ura,” a blend of artificial intelligence and aura, redefining artistic identity in the posthuman era.

Artistic Strategies Based on Transcoding

Mario Klingemann represents a generation of artists shaped by computational tools, yet his ambitions extended far beyond being a passive user of software. He quickly advanced to mastering scripting and programming, becoming not only the creator of his own tools but also a pioneer pushing the boundaries of artistic inquiry. This approach established the ontological foundation of his work, while expanding his exploration into dimensions particularly relevant in the posthuman era. This era redefines the creative process, where humans are no longer the sole agents, and artistic works emerge from the collaboration of human and non-human actors.

Klingemann’s strategies investigate the potential of artificial intelligence to analyze vast datasets of artistic styles, themes, and techniques across history while testing its capacity to produce works that both integrate and reinterpret these influences. He argues that art history itself becomes a key actor in this process. In today’s technological age, where the entirety of art history is figuratively “at our fingertips” through digital tools and algorithms capable of extracting its traits, ideas, and concepts, new possibilities for artistic practice emerge. This enables artists to not only reproduce past works but also use historical elements to create new artistic expressions or redefine what art can represent and become. Klingemann’s work exemplifies both technological sophistication and a critical reflection on the dynamic relationship between the past and the future of art.[18]

One of his pivotal works, The Butcher’s Son (2017), illustrates the synergy between machine learning and human intuition. Referencing the works of British painter Francis Bacon, the piece transcends simple imitation of Bacon’s aesthetic. Instead, it critically examines the process of training machine models on thousands of pornographic images, raising questions about dataset curation, attention, and emotional detachment. The Butcher’s Son is not a direct reproduction of any particular artist, style, or technique but the outcome of a machine learning process that mimics the natural creativity of artists who historically synthesized influences into original and provocative works. Through this lens, Klingemann reflects on artificial intelligence as a creative agent, akin to human artists, using historical materials to transform them into new conceptual and aesthetic forms. As Klingemann succinctly puts it, “To use history as clay to mold.”

This process also highlights questions about the limits and possibilities of artificial intelligence as a tool to transcend conventional boundaries in artistic discourse. Klingemann’s projects align with Lev Manovich’s concept of “deep remixability,” a principle introduced in After Effects, or Velvet Revolution (2006), which denotes the unprecedented ability to manipulate artistic elements — styles, epochs, media, and genres — based on their numerical essence.[19] By eliminating hierarchical relationships between genres and historical barriers between epochs, this approach allows combinations and transformations previously unimaginable in traditional art, expanding the scope of artistic discourse.

Klingemann’s practice also explores the epistemological implications of the numerical foundation of digital media. By leveraging transcoding — the translation of all languages, including visual ones — Klingemann enables meaningful interactions between human creativity and machine capabilities, creating a new paradigm of authenticity and authorship. His interactive installation Circuit Training (2019), presented at the More Than Human exhibition in the Barbican, invited visitors to participate in training a neural network to produce art. Through this process, the machine continuously learned from human interactions, evolving into a living piece of artistic creation.[20]

By integrating these strategies, Klingemann investigates the relationship between AI learning processes and the expectations that machines may have of their creators. While artificial intelligence lacks conscious intentionality or autonomous thought, its programming can incorporate human-defined goals and preferences. This “constructed intentionality” enables AI systems to act as though they have interests, even though these are merely reflections of human intentions embedded in their design. Klingemann’s work underscores the reciprocal relationship between humans and machines, where human decisions not only shape AI behavior but also transform how people perceive themselves and their place in a technologically interconnected world.

A(I)Aura Strategy

In his essay The Work of Art in the Age of Mechanical Reproduction, Walter Benjamin analyzes the profound impact of technological advancement on the creation and perception of art. His argument focuses on the concept of “aura” — the unique quality of artworks closely tied to their originality and authenticity.[21] According to Benjamin, mechanical reproduction disrupts this aura by enabling mass production and wide distribution, fundamentally altering how art is perceived and valued. This framework has been applied to various reproduction technologies of the industrial and modern eras, from photography to technical drawings and reprographic processes. However, in the context of AIArt, there is a significant shift away from the principles of mechanical reproduction. AIArt does not operate by merely copying or reproducing existing works but generates novel outputs through algorithms that reinterpret and transform historical styles, motifs, techniques, and even scientific methods.[22]

This process expands the boundaries of creativity while raising new questions about authenticity, originality, and the value of art in the digital age. Benjamin’s analysis offers a theoretical foundation for understanding shifts in art perception in the era of artificial intelligence, yet contemporary technological approaches require further development of his concepts to address the unique characteristics and implications of art-making in a post-digital environment.

The question of authenticity within the latent space or in AIArt challenges traditional notions tied to human perception of this phenomenon. It prompts a discussion on whether AIArt represents a new form of authenticity reflecting the distinct features of a technologically driven era, where human creativity synergizes with machine capabilities. In this context, authenticity can be seen as the capacity of algorithms to reinterpret or generate digital-cultural artifacts, producing a unique digital aura. Central to this discourse is the notion of the latent space — a multidimensional space where data is represented in compact and meaningful ways, allowing exploration of relationships and interpolation between data points. Interpolation within the latent space provides tools for meaningful manipulation and transformation of data, but only if there is a clearly defined goal.

Without such orientation, this movement risks resembling random wandering, raising further questions about AI’s ability to achieve deeper and meaningful artistic authenticity.[23]

Klingemann addresses this issue by distinguishing between strategies based on “closed systems” and “open systems.” In a closed system, the artist would construct a dataset consisting solely of previously created manual artworks and train the algorithm accordingly. This approach foregrounds the artist’s deliberate control over the dataset and training process.

However, Klingemann emphasizes the necessity of considering external factors that influence the datasets, including contributions from other artists and the historical context, whether personal or universal. In this sense, authorship becomes an illusion, regardless of whether the artistic creation system is closed or open. Open systems, such as popular image generators based on diffusion models, present significant challenges. According to Klingemann, companies developing these tools often intervene in the collective cultural memory by censoring dataset content and restricting the language used to generate inputs. He characterizes this approach as a form of “cultural memory lobotomy,” which reduces the availability and diversity of historical and artistic heritage, limiting its potential for new creations. This critique highlights the broader implications of technological mediation in shaping both cultural narratives and artistic practice.[24]

Conclusion

This study analyzes the artistic strategies of Mario Klingemann from 2018 to 2024. The research is based on a qualitative analysis of interviews with the artist, providing an authentic insight into his creative processes and conceptual foundations. The primary goal is to identify the methodological principles and epistemological implications of his work within the context of the posthuman era and the evolving relationship between humans and artificial intelligence.

Klingemann’s artistic projects reflect the concept of distributed agency, in which the act of creation is not solely attributed to the human subject but emerges as a synergy between human and non-human agents. This concept, rooted in theoretical discussions in philosophy of technology, posthumanism, artificial intelligence, and social sciences, challenges the traditional anthropocentric perception of agency as an exclusively human domain. Instead, it redefines agency as a decentralized and distributed process, in which the ability to act is shared among various human and non-human entities. In this context, the study examines the process of transcoding as a principle that enables the reinterpretation of historical artistic patterns through generative algorithms. This approach leads to a redefinition of authorship and authenticity in artistic practice, positioning the artwork as an emergent outcome of an iterative collaboration between humans and AI systems.

Another key theme of this research is the question of authenticity and originality in the digital environment. The study analyzes Walter Benjamin’s concept of the “aura” and its application in the context of AIArt, identifying a shift from static artistic objects to continuously evolving visual structures generated by artificial intelligence. This shift leads to the formulation of the concept of A(I)aura, which represents a new form of authenticity emerging within AIArt — where the artwork acquires a fluid and process-oriented nature.

The study’s findings suggest that Mario Klingemann’s artistic strategies make a significant contribution to contemporary discussions on the role of the artist in the era of artificial intelligence. These strategies not only redefine the relationship between the artist and the medium but also open broader philosophical and epistemological questions regarding the future of artistic creation in a technologically interconnected world. In this way, the study contributes to a deeper understanding of artistic research in the field of AIArt while providing a theoretical framework for further exploration in this area.

Additionally, the findings indicate a potential connection between artistic strategies and the discourse on consciousness in artificial intelligence systems, referencing the study Consciousness in Artificial Intelligence (2023). This research provides methodological frameworks for examining agency and cognitive processes within AI systems, creating opportunities for further interdisciplinary analyses. Klingemann’s work reflects these insights through the concept of interactive AIArt, where artificial intelligence is not merely a tool but an active creative partner. As a result, the artistic production process becomes a dynamic network of relationships between algorithms, the artist, and the audience, leading to a fundamental redefinition of authorship, originality, and the perception of the artwork in the digital age.

References:

ARCHER, Louise, Emily Dawson, Jennifer DeWitt, Amy Seakins, and Billy Wong. 2015. “Science Capital: A Conceptual, Methodological, and Empirical Argument for Extending Bourdieusian Notions of Capital Beyond the Arts.” Journal of Research in Science Teaching, vol. 52, no. 7, pp. 922–948. DOI: 10.1002/tea.21227.

BARTHES, Roland. 1967. “The Death of the Author.” Aspen, no. 5–6. Available online: http://www.ubu.com/aspen/aspen5and6/threeEssays.html.

BECKER, Howard S. 1982. Art Worlds [online]. Berkeley, CA: University of California Press. ISBN 9780520052185. Available from: https://books.google.sk/books?id=jXDyRK2EL5YC.

BENJAMIN, Walter. 2008. The Work of Art in the Age of Mechanical Reproduction. Harlow: Penguin Books.

DEL CAMPO, Matias. 2024. “Art Beyond Mechanical Reproduction: In Conversation with AI Artist Mario Klingemann.” Architectural Design, vol. 94, no. 3, p. 64. DOI: 10.1002/ad.3056.

DPRIX, Wolf, et al. 2022. “The Legacy Sketch Machine: From Artificial to Architectural Intelligence.” In LEACH, Neil, and Matias Del Campo (eds.), Machine Hallucinations, no. 3, pp. 14–21.

FROW, John, and Michael Emmison. 1998. “Information Technology as Cultural Capital.” Australian Universities’ Review, vol. 41, no. 1, pp. 41–45.

GIANNACHI, Gabriella. “Framing Humans for AI.” In PERISSINOTTO, Luigi (ed.), JoLMA: The Journal for the Philosophy of Language, Mind and the Arts [online], p. 56. Available from: https://edizionicafoscari.unive.it/en/edizioni/riviste/the-journal-for-the-philosophy-of-language-mind-an/.

KLINGEMANN, Mario. 2018. Memories of Passersby I [online]. Available at: https://www.sothebys.com/en/articles/mario-klingemanns-pioneering-ai-artwork.

MANOVICH, Lev and Emanuele Arielli. 2019. Artificial Aesthetics: A Critical Guide to AI, Media, and Design. 2023. [S.l.]: Available at: http://manovich.net.

MANOVICH, Lev. 2001. The Language of New Media. Cambridge, MA: MIT Press.

MANOVICH, Lev. 2007. “After Effects, or Velvet Revolution.” In Artifact, vol. 1, no. 2, pp. 67–75.

MANOVICH, Lev. 2019. Defining AI Arts: Three Proposals [online]. Available at: http://manovich.net/content/04-projects/107-defining-ai-arts-three-proposals/manovich.defining-ai-arts.2019.pdf.

MANOVICH, Lev. 2024. Lev Manovich on LinkedIn: A Summary of My Recent Post “There Is No AI Aesthetics” in 7 Slides… [online]. May 2024. Available at: https://www.linkedin.com/posts/levmanovich_a-summary-of-my-recent-post-there-is-no-activity-7203005246913060865-cxjR.

MANOVICH, Lev. 2024. [Online]. Available from: https://www.linkedin.com/posts/levmanovich_a-summary-of-my-recent-post-there-is-no-activity-7203005246913060865-cxjR.

NELSON, Jason. 2023. “Midjourney’s AI Platform Blocks Images of Chinese President, Sparking Ethics Debate.” Decrypt [online]. Available from: https://decrypt.co/125345/midjourneys-ai-platform-blocks-images-of-chinese-president-sparking-ethics-debate.

RHEINBERGER, Hans-Jörg. 1997. Toward a History of Epistemic Things: Synthesizing Proteins in the Test Tube. Stanford, CA: Stanford University Press.

RHEINBERGER, Hans-Jörg. 2018. “Epistemic and Aesthetics of Experimentation: Towards a Hybrid Heuristics?” In Philippe Sormani, Guelfo Carbone, and Priska Gisler, eds., Practicing Art/Science: Experiments in an Emerging Field, 236–249. Abingdon, UK: Routledge.

RHEINBERGER, Hans-Jörg. 2023. Split and Splice: A Phenomenology of Experimentation. Chicago: University of Chicago Press.

SCHWAB, Michael, and Henk Borgdorff. 2014. The Exposition of Artistic Research: Publishing Art in Academia. Leiden: Leiden University Press. ISBN 9789087281649.

SCHWAB, Michael. 2018. “Transpositionality and Artistic Research.” In Michael Schwab, ed., Transpositions: Aesthetico-Epistemic Operators in Artistic Research [online], 191–214. Leuven: Leuven University Press. ISBN 9789462701410. Available at: http://www.jstor.org/stable/j.ctv4s7k96.14.

SCHWAB, Michael. 2023. “Research Exposition on Artistic Research and Transdisciplinarity.” HUB — Journal of Research in Art, Design and Society [online]. Vol. 0, No. HUB Special Issue — Issue #0 / Spring 2023. Available at: https://doi.org/10.22501/hub.2190234.

[1] MANOVICH, Lev. 2019. Defining AI Arts: Three Proposals [online]. Available at: http://manovich.net/content/04-projects/107-defining-ai-arts-three-proposals/manovich.defining-ai-arts.2019.pdf.

[2] MANOVICH, Lev and ARIELLI, Emanuele. 2019. Artificial Aesthetics: A Critical Guide to AI, Media, and Design. 2023. [S.l.]: Available at: http://manovich.net.

[3] MANOVICH, Lev. 2024. Lev Manovich on LinkedIn: A Summary of My Recent Post “There Is No AI Aesthetics” in 7 Slides… [online]. May 2024. Available at: https://www.linkedin.com/posts/levmanovich_a-summary-of-my-recent-post-there-is-no-activity-7203005246913060865-cxjR.

[4] SCHWAB, Michael. 2023. Research Exposition on Artistic Research and Transdisciplinarity. HUB — Journal of Research in Art, Design and Society [online]. 2023, Vol. 0, No. HUB Special Issue — Issue #0 / Spring 2023. Available at: https://doi.org/10.22501/hub.2190234.

[5] RHEINBERGER, Hans-Jörg. 1997. Toward a History of Epistemic Things: Synthesizing Proteins in the Test Tube. Stanford, CA: Stanford University Press.

[6] RHEINBERGER, Hans-Jörg. 2018. “Epistemic and Aesthetics of Experimentation: Towards a Hybrid Heuristics?” In Philippe Sormani, Guelfo Carbone, and Priska Gisler, eds., Practicing Art/Science: Experiments in an Emerging Field, 236–249. Abingdon, UK: Routledge

[7] RHEINBERGER, Hans-Jörg. 2023. Split and Splice: A Phenomenology of Experimentation. Chicago: University of Chicago Press.

[8] SCHWAB, Michael. 2018. “Transpositionality and Artistic Research.” In Michael Schwab, ed., Transpositions: Aesthetico-Epistemic Operators in Artistic Research [online], 191–214. Leuven: Leuven University Press. ISBN 9789462701410. Available at: http://www.jstor.org/stable/j.ctv4s7k96.14.

[9] SCHWAB, Michael, and Henk Borgdorff. 2014. The Exposition of Artistic Research: Publishing Art in Academia. Leiden: Leiden University Press. ISBN 9789087281649.

[10] WAELDER, Pau, 2020. Beyond Ganism: AI as Conceptual Art. CIAC MTL [online]. 2020. Dostupné na: http://ciac.ca/en/ai-ciac-mtl-03-01/

[11] KLINGEMANN, Mario. 2018. Memories of Passersby I [online]. Available at: https://www.sothebys.com/en/articles/mario-klingemanns-pioneering-ai-artwork.

[12] BARTHES, Roland. 1967. “The Death of the Author.” Aspen, no. 5–6. Available online: http://www.ubu.com/aspen/aspen5and6/threeEssays.html.

[13] MANOVICH, Lev. 2001. The Language of New Media. Cambridge, MA: MIT Press.

[14] MANOVICH, Lev. 2024. [Online]. Available from: https://www.linkedin.com/posts/levmanovich_a-summary-of-my-recent-post-there-is-no-activity-7203005246913060865-cxjR.

[15] FROW, John, and Michael Emmison. 1998. “Information Technology as Cultural Capital.” Australian Universities’ Review, vol. 41, no. 1, pp. 41–45.

[16] ARCHER, Louise, Emily Dawson, Jennifer DeWitt, Amy Seakins, and Billy Wong. 2015. “Science Capital: A Conceptual, Methodological, and Empirical Argument for Extending Bourdieusian Notions of Capital Beyond the Arts.” Journal of Research in Science Teaching, vol. 52, no. 7, pp. 922–948. DOI: 10.1002/tea.21227.

[17] BECKER, Howard S. 1982. Art Worlds [online]. Berkeley, CA: University of California Press. ISBN 9780520052185. Available from: https://books.google.sk/books?id=jXDyRK2EL5YC.

[18] DEL CAMPO, Matias. 2024. “Art Beyond Mechanical Reproduction: In Conversation with AI Artist Mario Klingemann.” Architectural Design, vol. 94, no. 3, p. 64. DOI: 10.1002/ad.3056.

[19] MANOVICH, Lev. 2007. “After Effects, or Velvet Revolution.” In Artifact, vol. 1, no. 2, pp. 67–75.

[20] GIANNACHI, Gabriella. “Framing Humans for AI.” In PERISSINOTTO, Luigi (ed.), JoLMA: The Journal for the Philosophy of Language, Mind and the Arts [online], p. 56. Available from: https://edizionicafoscari.unive.it/en/edizioni/riviste/the-journal-for-the-philosophy-of-language-mind-an/.

[21] BENJAMIN, Walter. 2008. The Work of Art in the Age of Mechanical Reproduction. Harlow: Penguin Books.

[22] DPRIX, Wolf, et al. 2022. “The Legacy Sketch Machine: From Artificial to Architectural Intelligence.” In LEACH, Neil, and Matias Del Campo (eds.), Machine Hallucinations, no. 3, pp. 14–21.

[23] DEL CAMPO, Matias. 2024. “Art Beyond Mechanical Reproduction: In Conversation with AI Artist Mario Klingemann.” Architectural Design, vol. 94, no. 3, p. 64. DOI: 10.1002/ad.3056.

[24] NELSON, Jason. 2023. “Midjourney’s AI Platform Blocks Images of Chinese President, Sparking Ethics Debate.” Decrypt [online]. Available from: https://decrypt.co/125345/midjourneys-ai-platform-blocks-images-of-chinese-president-sparking-ethics-debate.

List of Figures:

Figure 1
 Klingemann, M. (2010, November 2). Schönes aus Code [Digital artwork]. Decoded Conference 2010. Source: Decoded Conference.

Figure 2
 Klingemann, M. (2020, August 29). Circuit Training [Digital artwork]. Under Destruction. Source: Quasimondo.

Figure 3
 Klingemann, M. (2018). The Butcher’s Son [AI-generated artwork]. © Lumen Prize.