Artificial Intelligence, Pseudo-Futurism and the Future of Art

Replication in AI art does not attempt to fill in absences; it instead erodes presence. It is reduced to a simulated sense. Since art is referential in itself, it becomes the referent of the referent. This seemingly eternal deferral detracts from the absoluteness and integrity of art. Object integrity is closely linked to originality of art. It is the wholeness and cultural groundedness of art, what gives art, or an object, its cohesive identity, even in the face of replication. It answers the question of who, what, where and why the artwork was made.  

Art produced in the style of Salvador Dali’s ‘Woman with the Head of Roses’ does not become original through replication. Like the museum sketches of the Benin Bronzes, it merely gestures toward an absent referent—highlighting its own derivative nature rather than asserting autonomy. Before an argument can be made for or against replication, we must understand what constitutes originality in art.  

In 1919, Marcel Duchamp made ‘L.H.O.O.Q’ or ‘La Joconde’, modified postcard of Leonardo Da Vinci’s ‘Mona Lisa’ with a moustache drawn on. This act of replication does not hinge on the erasure or the non-substance of the subject, what it does is to build on the presence of the artwork, preserving the integrity of the object. While replicated or mass-produced art gives room for the original to be, AI art erases the original altogether. The integrity of the object is the substance of the object. Substance or non-substance can also be seen in the light of Walter Benjamin’s definition of aura in his book, The Work of Art in the Age of Mechanical Reproduction. For Benjamin, aura is the quality integral to an artwork that cannot be communicated through mechanical reproduction technique. More than this, it relies on the confidence in the initial. The spatio-historic relevance of an artwork ensures that the original, although replicated a thousand of times, remains the original. AI art is aura-less. This is dependent, of course, on the creator, its spatio-historic relevance and the concept of authenticity.   

The originality or fauxness of generative AI art is underscored by the absence of the pointers of aura, as posited by Walter Benjamin. If the argument against AI-created art is declared deficient by including generative AI, then attempting to define AI art becomes recursive. Over the years, different replicas of artworks have been made, including ‘L.H.O.O.Q (Shaved)’, a version of ‘L.H.O.O.Q.’ None of these replicas or reinterpretations replaced the original. They exist as they are, replicas and reinterpretation, nothing more.  

AI art, however, does the opposite. The replica assumes the position of the original. It is no news that artificial intelligence runs on a database of uploaded information, ranging from pictures to texts. It is no extraterrestrial entity. The dataset did not materialize from thin air. Artificial intelligence can therefore create nothing beyond a recombination of what already exists. What informs the rightness or wrongness of replication is more than the process. It is the final product. Making a case for the ‘originalization’ of replicas is akin to the death of object integrity. It is the height of super-modernity, where the simulated assumes the position of the real, as the French philosopher Jean Baudrillard warned. It should be seen as it is: non-substance. It extends this non-substance outward by crowning itself the initial just like the British Museum insists on its rights to keep the bronze heads. 

Another essential aspect is the ethicality of artificial intelligence. Though authors and artists in general have spoken up about the use of their creations in training AI models. It continues anyway. How does one trust it to churn out anything original? As long as the human element remains in AI, whatever it creates can be regarded as referential. AI companies like OpenAI, similar to the British Museum, make excuses for this use, rely on systems of justification (‘fair use,’ ‘preservation’) to defend extraction without consent. The digital form of colonialism might look different, but the mechanism is the same.   

Some have compared today’s scepticism about AI art to the art world’s early response to photography and the daguerreotype. This skepticism is echoed in Walter’s definition of aura. But both the daguerreotype and the painting or other art forms regale in their referential aspect. While a copy or reproduction is often materially present but referentially clear, AI-generated art is not. Photography is not painting, nor does it attempt to be. Where does this comparison leave AI art? Without the elements of aura, it appears unmoored. It floats and is atemporal, untethered to neither signs nor place.  

Also, continued clamour for machine-created art erases the human element in both art creation and machine invention. As Jean Baudrillard wrote in Simulacra and Simulation, technology is an extension of the body. The future it promises is not here yet. The machine is not sentient. It is heavily reliant on its human element: creativity, design and oversight. It is pseudo-futurism steeped in digital colonialism. It mines what already exists, removes attribution, flattens context, then presents the result as ‘innovative.’ It is not the future, but the re-packaged present. Digital colonialism and appropriation are aided by the false prophets of pseudo-futurism and post-humanism, the Silicon Valley myth-makers. While the Benin Bronzes remain unrepatriated, we are already witnessing a new form of cultural theft. 

There is no simple solution to this. Acknowledging the true referential nature of artificial intelligence and how it recombines data sets is not enough. Beyond this, efforts should be made to update data protection law in relation to AI, as the response has not been a positive one so far. We must all be reminded of this truth: the world has not, will not, outgrown artists and creatives⎈