The Obligation of Truth: An Investigation into the Function of Art from Plato to Artificial Intelligence

Must art express truth? This question, as simple in its formulation as it is profound in its implications, spans the entire history of Western thought. It presupposes an almost moral link between the work and a value—truth—that seems to belong more properly to philosophy or science. Answering this question requires a preliminary investigation into what is meant by “truth” in an artistic context, an intrinsically non-discursive field. The thesis defended is that art is not obligated to conform to a pre-established notion of truth; rather, its deepest intellectual value lies in its ability to actively question, reconfigure, and generate new forms of meaning and truth. Let’s explore this by analyzing the classical framework of the problem, outlining the three conceptions of truth—as correctness, plausibility, and metaphysical revelation—that emerge from the confrontation between Plato and Aristotle. Subsequently, we will evaluate these categories in dialogue with contemporary perspectives, particularly with the processual and “artificialist” vision of Alexander Manu. Finally, we will explore how the advent of generative artificial intelligence not only urgently re-proposes the ancient question but radically transforms it, forcing us to renegotiate the very nature of artistic truth and the role of the artist.

The Ancient Question: Truth as Correctness, Plausibility, and Revelation

The philosophical debate on truth in art has its roots in ancient Greece. The fundamental confrontation between Plato and Aristotle gave rise to three distinct concepts of truth applicable to the artistic field. Plato, in the Republic, expresses profound skepticism. His critique is based on a metaphysical hierarchy: art, being an imitation of sensible objects which are themselves imitations of the perfect Forms, is three times removed from essential truth. At best, a work can be judged for its correctness in representing appearance, but even in this, it often fails, proposing morally and epistemically false models. However, the Platonic critique can be read not as a simple condemnation, but as a philosophical “invitation.” Plato challenges art to justify its existence in a well-governed city, to demonstrate that it can be not only pleasant but also “beneficial,” that is, to possess an autonomous cognitive value that is not merely illusory. Aristotle, in the Poetics, implicitly accepts this invitation but shifts the plane of discourse. For him, the domain of art is not factual reality, but possibility. The poet should not describe “things as they have been,” but “things as they could be.” Consequently, the criterion for evaluation is not correctness, but aesthetic plausibility. A work is artistically valid if its plot is coherent and represents universals of human behavior in a convincing way. In this sense, poetry is “more philosophical and more serious than history.” Aristotle does not solve Plato’s metaphysical problem, but dissolves it, transforming it into an epistemic question: art does not reveal “being in itself,” but provides a form of knowledge and emotional understanding about the human world. From the classical world, we therefore inherit a complex conceptual grid: truth as factual correspondence (correctness), truth as aesthetic coherence (plausibility), and truth as revelation of essence (the still-open Platonic challenge).

The Contemporary Shift: Truth as an Artificial Process

Contemporary thought has further complicated this framework. Alexander Manu, in his essay Transcending Imagination, offers a radical perspective, perfectly suited to the digital age, starting from the assumption that “all art is artificial.” This statement dissolves the nostalgia for “natural” or mimetic art. Art is not a copy of nature, but a construction from its very beginning. This view breaks down the wall between human and machine creativity: if all art is artificial, then AI creations are simply another form of artifice, which fits into a long continuum. Manu shifts the focus from the final product to the process, which he breaks down into the triad of “Intention-Articulation-Manifestation.” Intention is the meaning, the concept. Articulation is the energy spent to translate it into a communicable form (for AI, the prompt). Manifestation is the finished work. In this new paradigm, the human artist is transformed from a “creator” to a “storyteller”: their skill no longer lies in the technical mastery of manifestation, which can now be delegated to the machine, but in the depth of their intention and the poetic richness of their articulation. The “truth” of a work, therefore, is not found only in the result, but in the coherence and power of the process that generated it, starting from human intention.

Artificial Intelligence: The Crisis and Generation of Truth

The advent of generative artificial intelligence forces a reckoning with this philosophical baggage. AI challenges our traditional notions of artistic truth, mastering some and exposing others. Truth as Correctness: AI can achieve a nearly perfect level of mimetic correctness, generating photorealistic images indistinguishable from reality. However, as 20th-century art history taught us after the invention of photography, mere technical accuracy rarely coincides with artistic value. Perfect correctness, if devoid of a deeper intention, risks producing a “sterile kitsch.” Truth as Plausibility: This is where AI excels. Having been trained on a vast archive of human works, it is capable of generating infinite plausible creations, consistent with the styles, genres, and archetypes it has learned. It can create a work “in the style of Van Gogh” with astonishing verisimilitude. However, as Manu warns, this virtuosity risks being a reproduction of existing archetypes, not the creation of authentically new forms. Truth as Revelation: This is where the decisive game is played. Can an AI respond to Plato’s invitation? Can it create a work that reveals a non-discursive truth, a meaning that transcends the statistical sum of its training data? Truly creative art, it has been argued, must represent something that does not yet “exist” in the known symbolic world. AI, by its nature, operates on what has already been said and seen. It seems condemned to remain in the Platonic domain of the imitation of imitations. And this is where the role of the human artist, as theorized by Manu, becomes crucial. If AI masters plausibility, the artist’s task is no longer to create plausible works, but to formulate articulations (prompts, systems, processes) so radical and unexpected that they force the machine to break out of its patterns, to generate that “incidental beauty,” a form of emergent revelation that was neither in the intention of the human nor in the machine’s schemes.

Conclusion

To the question, “must art express truth?” the answer that emerges from this journey is a qualified no. Art does not have to conform to a single, rigid notion of truth. On the contrary, its most vital function is to be a perennial laboratory where our conceptions of truth are tested, hybridized, and generated ex novo. Art is not obligated to be true in the sense of being factually correct or even just plausible. These are domains that artificial intelligence is rapidly conquering. The true vocation of art, its response to the Platonic challenge, lies in its ability to generate a truth as revelation: a constellation of meaning that cannot be expressed otherwise, a form that does not illustrate pre-existing knowledge but creates new knowledge. In the age of AI, this function becomes more critical than ever. The artist is not the one who creates a beautiful image, but the one who, through a profound intention and masterful articulation, questions the machine in such a way as to bring out the unexpected. The truth of art will not reside in the AI’s answer, but in the human’s question.

Essential Bibliography

• Boris Groys, The Truth of Art, e-flux journal, no. 71, 2016. A crucial essay for understanding how the notion of truth is transformed in contemporary artistic production and circulation.

• Errol Boon, The Question of Artistic Truth: Preparations for a Critique of Non-Discursive Reason (Master of Arts thesis, 2023). A study that distinguishes between factual correctness and aesthetic plausibility, and supports the idea that art contributes to understanding the world.

• Nelson Goodman, Languages of Art (1968). A classic of analytic philosophy of art: introduces the concept of “worldmaking” to explain the cognitive role of artistic representations.

• Noël Carroll, Philosophy of Art: A Contemporary Introduction (2000). A clear introduction to the philosophy of art, useful for contextualizing the debate on artistic truth.

• Catherine Z. Elgin, Considered Judgment (1996). The author argues that metaphors, fiction, and works of art also contribute to knowledge, despite not being literally true.

• Plato, Republic, books II and X. Fundamental for understanding the original distrust of art as a deceptive imitation.

• Aristotle, Poetics. Foundational text for the theory of verisimilitude as a cognitive path through art.

• Walter Benjamin, The Work of Art in the Age of Mechanical Reproduction (1936). A reflection on the change in the value (and truth) of art in the age of technology and media.

• Hans-Georg Gadamer, Truth and Method (1960). Proposes an idea of hermeneutic truth: the deep understanding of meaning, not scientific verification.

• David Freedberg and Vittorio Gallese, Motion, Emotion and Empathy in Esthetic Experience, Trends in Cognitive Sciences, 2007. An interdisciplinary study on how artistic perception activates cognitive and affective mechanisms that lead to a form of embodied understanding.