Bernabé Mallo
Doctor en Filosofía por la Universidad del País Vasco (UPV/EHU)
Investigador
en neurofilosofía, evolución humana y origen del arte. / PhD in
Philosophy – University of the Basque Country (UPV/EHU)
Researcher in neurophilosophy, human evolution, and the origins of art.
A review of Leonardo Arriagada's article (2025): Defining an AI-Generated Artwork: A Transdisciplinary Concept for Cognitive Science, Computer Science, and Art Theory
Introduction: the Tower of Babel of artificial intelligence
An AI-generated artwork is, for a computer scientist, a set of probabilistic outputs from a neural network. For a cognitive psychologist, a stimulus that provokes certain responses in the viewer. For an art theorist, an object that challenges traditional categories of authorship, intentionality, and creativity. Each discipline speaks its own language, and the result is often a Tower of Babel where communication is difficult and collaboration nearly impossible.
Leonardo Arriagada, in an article published in 2025 in the journal Calle 14: Revista de Investigación en el Campo del Arte, addresses precisely this problem . His goal is ambitious: to develop a transdisciplinary definition of an AI-generated artwork that serves as a conceptual bridge between cognitive science, computer science, and art theory.
Arriagada's proposal is structured around three essential elements: (1) the autonomous production by AI of a new and surprising idea or artifact; (2) the existence of an internal evaluation mechanism within the system itself; and (3) the consideration of that product as a candidate for appreciation by a human audience.
This definition, as we shall see, is not without philosophical problems, but it has the merit of offering a common starting point for disciplines that often speak without understanding each other. From the perspective of our research on the S/Y/C model and Surgical Philosophy, Arriagada's analysis is especially valuable because it forces us to clarify what we mean by "autonomy", "evaluation", and "appreciation" in the context of machine-generated art.
Why do we need a transdisciplinary definition?
Arriagada starts from a realization: the literature on AI-generated art is abundant but fragmented . Computer scientists focus on the technical aspects of generative algorithms (generative adversarial networks, transformers, diffusion models). Cognitive psychologists study how viewers perceive and value these works. Art theorists debate whether AI can be considered an author or whether its products deserve the status of art.
The problem is not merely academic. It has practical implications for copyright attribution, critical evaluation, conservation, and commercialization of these works. Without a common language, disagreements are inevitable and opportunities for interdisciplinary collaboration are lost.
Arriagada's proposal does not pretend to be the definitive definition, but rather a heuristic tool that allows researchers from different fields to dialogue on common ground. His approach is pragmatic: he does not ask what art is in essence, but rather what characteristics an AI-generated object must meet to be considered a legitimate candidate for artistic appreciation.
The three elements of the definition
Arriagada proposes that an AI-generated artwork must meet three conditions .
1. Autonomous production of novelty
The first element is that the AI must produce something new and surprising autonomously. This excludes cases where the system merely reproduces or combines existing patterns without generating true novelty. It also excludes cases where human control is so direct that the AI acts as a mere tool, not as an agent.
The notion of "autonomy" is key here, but also problematic. How autonomous must the AI be? Is it enough that there is no human intervention at the moment of generation, or must the system have learned its criteria on its own without supervision? Arriagada does not offer a definitive answer, but notes that the degree of autonomy can vary and that this variability should be the subject of empirical study.
2. Internal evaluation mechanism
The second element is that the AI must incorporate an internal evaluation mechanism. It is not enough for it to generate something new; the system must be able to evaluate its own productions, to discriminate between successful and unsuccessful results, to select those that deserve to be presented to the human audience.
This mechanism can be explicit (for example, a loss function that penalizes certain outputs) or implicit (for example, the model's own statistics). But without some kind of internal evaluation, the generation would be blind, and the result a mere product of chance.
The inclusion of this element is interesting because it introduces a form of weak intentionality in AI. The system has no consciousness or purposes, but its architecture incorporates value criteria that guide its production. This brings it, however vaguely, closer to the notion of "judgment" that in humans accompanies creativity.
3. Candidate for human appreciation
The third element is that the AI-generated product must be considered a candidate for appreciation by a human audience. Here Arriagada introduces a social and institutional factor. It is not enough for the AI to produce something novel and evaluate it internally. That something must be presented to humans who can appreciate it, interpret it, value it.
This element recognizes that art is, in part, a social institution. An AI-generated work that remains on a computer's hard drive without anyone seeing it is not art, just as Duchamp's readymade would not be art if no one had exhibited it in a gallery. Human appreciation is the act that completes the circuit, that turns a technical artifact into an aesthetic object.
Strengths and weaknesses of the proposal
Arriagada's definition has several strengths. First, it is operational: it provides criteria that can be applied empirically to classify specific cases. Second, it is inclusive: it does not a priori exclude any type of technique or style, as long as the three conditions are met. Third, it is transdisciplinary: it offers a common language for computer scientists, psychologists, and art theorists.
However, it also has important weaknesses. The most obvious is that the definition depends on notions ("novelty", "surprise", "autonomy", "appreciation") that are themselves controversial and require further precision. What counts as "surprise" for a machine? How do we measure "autonomy"? Who decides what is "worthy of appreciation"?
Furthermore, the definition does not directly address the question of intentionality in the strong philosophical sense. A system can have an internal evaluation mechanism without there being anyone "inside" who actually judges or values anything. Evaluation is a function, not an experience. And that distance between function and experience is, as we have seen in other reviews, crucial for distinguishing human art from machine art.
Connection with research on the origin of art (S/Y/C)
Arriagada's analysis resonates deeply with the research we have been developing on the S/Y/C model of neuronal functioning and the Law of Biological Coherence. His transdisciplinary definition can be reinterpreted in light of our three dimensions.
The S (Survival) dimension reminds us that human appreciation of art is not a purely cognitive act, but is rooted in homeostatic needs. When a human viewer "appreciates" an AI-generated work, their brain is processing stimuli that generate pleasure, activation, curiosity, or emotion. That processing is real, even if the work was made by a machine. Arriagada's definition, by including "human appreciation", implicitly recognizes this dimension.
The Y (Symbolon) dimension is central to understanding the gap between AI's internal evaluation and human appreciation. AI can have evaluation mechanisms that are mathematical functions, but not symbolon: there is no shared recognition, no emergent meaning from an encounter between subjectivities. Human appreciation, by contrast, is always symbolic: we see the work, interpret it, connect it with our personal and cultural history. That is something that AI, by itself, cannot generate. Symbolon —the act of recognition through shared codes— is the patrimony of embodied subjectivity.
The C (Wholeness) dimension points to the human need to close forms, to find coherence. When we appreciate a work, we seek in it a totality that satisfies us. AI can generate formally closed objects, but it does not experience the drive toward wholeness. Its "internal evaluation" is a calculation, not a longing.
Surgical Philosophy invites us to make a precise analytical cut in Arriagada's definition. It is not about rejecting it, but about distinguishing levels. At the level of technical production, the definition is useful: it allows us to identify which AI-generated objects deserve to be considered candidates for artistic appreciation. At the level of full aesthetic experience, however, the definition is insufficient, because it does not capture what human appreciation has that is embodied, symbolic, and yearning.
Arriagada offers a tool for building bridges between disciplines. That is valuable. But to understand art in its origin and its function, we need something more: a theory of embodied subjectivity, of lived intentionality, of the drive toward meaning. That is what our research attempts to contribute, with symbolon as a central concept.
Implications for future research
Arriagada's article opens several lines for future research . First, it suggests that the proposed definition should be applied to various types of AI-generated artworks to assess its scope and limits. An image generated by a neural network is not the same as a poetic text produced by a language model, or a symphony composed by an autonomous system. The definition should be able to encompass them all, or be adjusted to do so.
Second, Arriagada notes that the definition has implications for artistic practices . If human artists begin to collaborate with AI systems that meet the three conditions, how does their way of working change? Do they remain the authors, or is authorship distributed? What happens to the notion of "style" or "voice" when the machine can imitate any style?
These questions have no easy answers, and Arriagada's definition does not pretend to resolve them. His contribution is more modest but equally valuable: to provide a common ground for researchers from different disciplines to begin a dialogue. And that dialogue, as we know, is the condition of possibility for any significant advance.
Final considerations: the need for a common language
Leonardo Arriagada's article reminds us that, in such a novel and controversial field as AI-generated art, the first challenge is not to answer questions, but to formulate them well. And to formulate them well, we need clear and shared concepts. His proposal for a transdisciplinary definition is a step in that direction.
From our perspective, that definition is useful but incomplete. Useful because it allows communication between cognitive science, computer science, and art theory. Incomplete because it does not capture the embodied dimension of human aesthetic experience. AI can produce objects that meet Arriagada's three criteria, but those objects, when appreciated by a human, acquire a meaning that is not in the machine, but in the encounter between the work and a living subjectivity.
The origin of art, in the Homo species, was not the production of novel objects or the internal evaluation of autonomous systems. It was the need to express, to communicate, to share what one feels —it was the need for symbolon. That need remains, today, what defines the human against the artificial. And as long as we do not forget that, we can use AI as a tool without confusing it with the source.
References
Arriagada, L. (2025). Defining an AI-generated artwork: A transdisciplinary concept for cognitive science, computer science, and art theory. Calle 14: Revista de Investigación en el Campo del Arte, 20(38), 95–109. https://doi.org/10.14483/21450706.21009
Mallo, B. (2023). La construcción neuro-simbólica. Una aproximación al funcionamiento del cerebro desde una perspectiva multidisciplinar [Doctoral thesis, University of the Basque Country - Euskal Herriko Unibertsitatea]. ADDI Repository. http://hdl.handle.net/10810/62701
Mallo, B. (2026a). De la filosofía quirúrgica a la ley de coherencia biológica S/Y/C: Hacia una investigación sobre el origen del arte en la especie Homo. Lopez Mallo, Javier Bernabé. https://isbn.bibna.gub.uy/catalogo.php?mode=detalle&nt=57196
Mallo, B. (2026a). De la filosofía quirúrgica a la ley de coherencia biológica S/Y/C: Hacia una investigación sobre el origen del arte en la especie Homo [Kindle edition]. Amazon. https://www.amazon.com/dp/B0GYGTJD5C
Mallo, B. (2026b). From surgical philosophy to the law of biological coherence S/Y/C: Toward a study of the origin of art in the Homo lineage. Lopez Mallo, Javier Bernabé. https://isbn.bibna.gub.uy/catalogo.php?mode=detalle&nt=57197
Mallo, B. (2026b). From surgical philosophy to the law of biological coherence S/Y/C: Toward a study of the origin of art in the Homo lineage [Kindle edition]. Amazon. https://www.amazon.com/dp/B0GY89SZS1
Autor / Author
Bernabé Mallo
Doctor en Filosofía – Universidad del País Vasco / Euskal Herriko Unibertsitatea (UPV/EHU)
Investigador independiente en neurofilosofía, evolución humana y origen del arte.
Bernabé Mallo
PhD in Philosophy – University of the Basque Country / Euskal Herriko Unibertsitatea (UPV/EHU)
Independent researcher in neurophilosophy, human evolution, and the origin of art.
Enlaces / Links
Página de autor Amazon / Amazon Author Page: https://www.amazon.com/author/bernabemallo
ORCID: https://orcid.org/0000-0001-9002-9728
Plataforma EHUenRed / Link EHUenRed: https://www.ehu.eus/es/web/masterrak-eta-graduondokoak/red-latinoamericana-de-posgrados
Canal YouTube / Channel YouTube: https://www.youtube.com/@neuroideas815
Canal YouTube / Channel YouTube: https://www.youtube.com/channel/UCBsf6OZ482NjST6QA-hvYtQ
Publicaciones y proyectos en desarrollo / Publications and projects:
https://www.amazon.com/author/bernabemallo
https://ehuenred.theglocal.network/ideas/el-origen-del-arte-en-el-cerebro-de-makapansgat-al-moma-del-primate-al-sapiens