With the ever-heightening presence of AI, I find myself floundering, entrapped by the fear of being replaced, both physically and metaphysically, by this luridly humanlike intelligence and the countless shape-shifting cousins that exist under its sprawling wings such as ChatGPT, BlenderBot, Rasa, Stable Diffusion, and Midjourney. Physically, because AI can now drive for me, cook for me, clean for me, and play video games for (and with) me. Metaphysically, because AI can now assemble abstract paintings, compose poems, expound philosophy, and dissect art on my demand – all it needs is a glass of red wine and a few cigars to metamorphose into da Vinci or Schopenhauer!
All of this necessarily and invariably leads to my question: Can AI create art? Well, yes – if by art we mean objects that elicit aesthetic pleasure. DALL-E or Midjourney can produce visual pieces with a speck of the time we need and twenty times the symmetry, precision, colorational congruence, and so on. Nevertheless, I find myself dissatisfied with this answer. Art, in my opinion, transcends a passive object that evokes emotions from the viewers. Instead, it is a dialectical dialogue between the artist and their consciousness during which they draw upon their subjective experiences to connect to the subjective experiences of their audience.
To begin, let us consider a hypothetical scenario posed by American philosopher and mathematician Hilary Putnam. An ant is crawling on a patch of sand. As it crawls, it traces a line. By sheer chance, the line crisscrosses and zigzags itself in such a way that it resembles a recognizable caricature of Winston Churchill. Putnam argues that, given the absence of direct exposure to Churchill and visual representations thereof, the ant, by virtue of its limited perceptual capacities, couldn’t have harbored the intention to depict him. Instead, the ant engages in a purely mimetic act which, by happenstance, our cognitive faculties as observers interpret as a likeness of the former prime minister of the United Kingdom. In fact, even if Churchill did not exist, the ant still would have traced the same lines. We can see, therefore, that while a viewer may interpret a piece as art, the absence of intention on the creator’s part makes the production fundamentally futile.
Intention, however, is precisely what AI lacks. Like the ant, AI generates pieces by a string of probability calculus and keyword associations. ChatGPT, for example, adopts a transformer neural network to create text-based content in response to input prompts. When provided with a command such as “Compose a rap battle between Einstein and a mischievous quantum particle” or “Create a tale about emojis revolting and finding their missing punctuation partners,” GPT tokenizes this text, converts it into numerical representations and generates an output based on learned relational patterns from its training data. Similarly, DALL-E 2 utilizes a similar approach to construct visual pieces pixel by pixel. Devoid of intention and sensory experiences, AI lacks the subjective consciousness to genuinely recognize, identify, and appraise the content it creates.
Subjective memory also plays an essential role in art-making. When producing an artwork, the artist imbues their work with a meaning surpassing the sum of its parts. This process probes the nebulous depths of the artist’s memories, unveiling the consciousness, sub-consciousness, disorders, and illogicality only accessible to the human reflective mind. In his novel In Search of a Lost Time, Marcel Proust aptly illustrates the profundity to which an artist’s past experiences shape their craft. Through acts of what he terms “intense remembering,” our subjective formative histories may be transmuted into artistic representations. Rendered with singular perception, memories are resurrected and reshaped through mental alchemy in which personal reminiscence burgeons from the artist’s subjective lens. The language, signs, and symbols used in the art are interdependent components that derive their impact and substance through the distinctive perspectives of the artist. It is this subjective vision embedded in the creator’s mind that breathes life into their creation and grants art its capacity to reform. Art, thus, seeps past technicalities and aesthetic preferences by embodying the artist’s essence sculpted by their subjective experiences. This linkage between an artist’s interiority and their creation distinguishes true art from mere replications of form and style via computerized assemblages. AI cannot produce art as it lacks the lived experiences situated in space and time that inform an artist’s interiority and suffuse their creations with subjective qualities distinctive to the human condition. This concept loosely connects to philosopher John Searle’s well-known Chinese Room Theory, which argues even if an AI machine can speak perfect Chinese, it cannot truly understand the underlying semantics and meaning of this language. AI, thus, can emulate human neurological processes but ultimately falls short of comprehending the meaning of what it outputs. Put differently, they can only adhere to a strictly predetermined algorithmic procedure that is in itself incapable of thinking and comprehending.
In Heidegger’s philosophy, thinking is a continual activity that necessitates profound engagement with one’s existence and the surrounding world. It is grounded in the notion of “Dasein,” loosely translated from German as “being-there.” Dasein encompasses the complete awareness inherent in the human experience of existing and engaging with their surroundings. Heidegger posits that genuine thinking entails an ongoing and participatory inquiry into our existence, a process simultaneously private and universal. Now, the view that the true criterion of sentience is availability to consciousness is based upon the idea of genuine mental states. Processes investigated by cognitive science while unavailable to consciousness fail to qualify as genuine mental states – lacking consciousness, AI becomes one of such unqualified candidates. While AI is capable of content generation, it does not exercise human-level thinking and Dasein which enables the conscious appointment of meaning and intent to its creations the way humans do. This also connects to computer scientist Douglas Hofstadter’s criticism of machine translation on its absence of fundamental understanding. “It’s all about ultra-rapid processing of pieces of text, not about thinking or imagining or remembering or understanding” that is at once “nonverbal, imagistic, experiential, mental”, something AI innately lacks.
We also see this concept of subjective consciousness embodied in art. In Johannes Vermeer’s Woman in Blue Reading a Letter, we see a young lady clutching onto a slice of paper, mouth half-opened, eyes transfixed on the content of the letter which we, the audience, can never know. The title of the artwork reinforces this sense of self-containment and inscrutability, its depiction no more than an objective repetition of the observable aspects of the artwork. The thoughts, emotions and inner worlds of this woman remain private and exclusively hers, veiled in a sort of absorbed silence that is at once resplendent and hushed. Vermeer masterfully arrested his subjects’ most intimate moments in his work, their inner worlds tantalizingly elusive and wholly self-contained. Vermeer gave life to this woman through his own subjective consciousness and interpretations. We then proceed to interact with the work by injecting our own experiences into our decoding of the work to synthesize a new meaning shared exclusively between us and the artist.
Similarly, in another work by him, A Maid Asleep, Vermeer depicted the thoughts of the young girl suspended in a space-time lacuna. What is she thinking? What thoughts haunt her mind? We may never know the real answer. What we do know, however, is that by fusing our subjective thoughts with that of the artist, we connect as humans. It is through this very inaccessibility of thought that Vermeer paints the essence of art as a patchwork of emotions and consciousnesses at once splendorously self-contained and universal on a human level.
An artwork stems from the artist’s individual interior experience draped within a specific spatial and temporal locality in their personal narrative. To define art otherwise would fail to recognize the essential role of human subjectivity in its creation and purpose-giving. Art requires human intention, the desire to convey private experiences and consciousness through visionary representations. Put differently, if art were to represent anything beautiful or aesthetically pleasing, then we may readily call the accidental arrangement of kitchen utensils in a sink or the haphazard scattering of toys on a living room carpet the epitome of artistic brilliance, equivocating the deliberate thought, intention, and craftsmanship artists infuse in their work with arbitrary happenstance. My view of art necessitates a human standpoint derived from an artist’s memories, sentiments and life experiences situated within a particular context, place, and era. To plead the case for nature and probability alone in the production of art would fail to appreciate what defines human subjectivity and the creative yields that stem from it.
By recognizing AI creations’ lack of human subjectivity in true art, we can more deftly leverage it as a means to complement and refine human creativity as opposed to replacing it. AI is the hammer to the blacksmith, the chisel to the sculptor. Rather than art theft, AI can function as a powerful tool with which artists express, define, and explore themselves. AI can aid as the brush and the paint, the mechanical means with which artists accomplish art whose legitimacy resides in its creator’s interiority and intention. Its purpose is to assist, not assert; to foster, not fetter; to refine, not define; to support, not supplant.