The contradiction of an inhuman art

In today's world, there is no need for fiction: reality is more absurd, engaging and full of twists and turns than anything George R.R. Martin could ever write. Yet there is an interesting phenomenon taking hold online: it is called, for lack of a better term, "unreality," and it began with a film called Goncharov, called "the greatest mafia movie ever made." The film, directed by Martin Scorsese in 1973, stars Robert De Niro as a Russian gangster who moves undercover to Naples and becomes a local mob boss. There's just one problem: the film does not exist and has been collectively created by different Tumblr and Twitter users who have been gradually imagining its existence, designing a poster, establishing a cast, editing GIFs depicting its scenes, and even writing a plot that includes secondary characters and love triangles. In November, a fan also composed a soundtrack for the film whose fame grew first when Neil Gaiman mentioned it as «Tumblr inside joke», then it became the subject of an article in The New York Times and finally, as a final triumph, it found joking confirmation with Martin Scorsese himself who, informed by his daughter Francesca of the phenomenon, claimed to have actually directed it. The thing did not end there. Taking advantage of the artificial intelligence MidJourney, in recent months, web users around the world have begun creating scenes and posters of films that never existed: Höllenhund, a horror film directed by Fritz Lang in 1919 and lost in a fire; Transformers if it was German Expressionist film; a 1950s adaptment of The Lord of the Rings directed by Ed Wood; Star Wars if it was a kabuki play; versions of Batman, The Avengers or The Shining directed by Wes Anderson, Tron in Alejandro Jodorowsky's style, The Grinch in Kubrick's or The Matrix if it had been filmed in the 1980s by Ridley Scott with sets designed by H. R. Giger.

Goncharov's game stimulates a new form of creative production, namely the operation (very Borgesian actually) of creating by induction something real from a few imaginary details. Yet this cinephile uchrony powered by the MidJourney software ends up involving the difficult question of human creativity in the world of artificial intelligences. Let's explain: many of today's creatives, from designers/painters to writers and translators, say they are alarmed by the rise of AIs capable of writing and illustrating books, translating texts or generating images that would require months of study and work for a normal human being. In the same days that AI art apps like Lensa are rising to the top of the App Store Top 10 for the first time in history, a controversy has erupted over ArtStation, perhaps the most relevant online platform in promoting independent graphic arts, which, as Kotaku explains «has no policy directly restricting the hosting or display of AI-generated imagery on the site, which has led to repeated instances where images made by computers, and not humans, have floated to the top of ArtStation’s “Explore” section». So the artists on the platform replaced their profile images with the classic ban sign, thus revealing which were human based and which were AI based. The struggle is not philosophical but economic and legal: the scraping technology by which artificial intelligences scan the Internet and retrieve the "atoms" used to generate their output is based on pre-existing images that include those of human artists, representing according to some a more elaborate form of plagiarism. Indeed, looking on Reddit at stills from a fictional Dario Argento film named Inferno di Stelle one can see how many elements are copied from 2001: A Space Odyssey, others taken from photographs of actress Barbara Steeele, and so on, all reshuffled to look like an original film. Artist Zed Edge summed it up on Twitter:

«When referecing art, humans use [references] as inspiration, AI uses them as samples. Without references, humans can still visualize ideas but AI cannot – both benefit from references but AI depends on them. AI doesn't "learn" from art it manipulates art beyond recognition, as a deception to appear new. Without consent or compensation to the original artist, AI art is theft».

Earlier this month the world of Unreality and the world of legally protected intellectual property faced a hitherto unsolved dilemma. Justin Roiland, creator of Rick & Morty, produced the video game High On Life, a sort of sci-fi/comedy adventure in whose gameplay are fictional movie posters created with MidJourney, describing the AI platform as «a tool that has the potential to make content creation incredibly accessible». Which put the issue of AI art on another level since in this case it's being used to create commercial assets instead of a salaried artist. To whose benefit, then, does creativity become more accessible? In fashion, too, the use of AI generates the same concerns: the Cala softwar, for example, can generate designs from pre-existing keywords or images. In the pages of BoF, the platform's CEO, Andrew Wyatt, argues that providing images or even just verbal descriptions to a design team, a practice typical of Virgil Abloh, Miuccia Prada, and Rei Kawakubo, none of whom draw sketches or create prototypes, is not too different from entering a prompt into an AI-managed image generator, which, by the way, are programmed not to copy their "cues" too closely - nor, the author of the piece argues, is the idea of a designer copying something never seen or heard of in fashion. We are still far from a takeover of artificial intelligences, no less, the existence of such platforms in fashion hints at a future where AIs may design garments based on an algorithm programmed to intercept and learn trends. Which would be worrisome, although Wyatt compares a designer using AI to a mathematician using a calculator.  Similarly, proponents of AI in video game design cite the speed and gratuitousness by which complex gameplays such as Red Dead Redemption, to take one example, can be animated without the manpower and staff that would normally be needed.

Those who support the use of AI in its diverse and endless applications argue that AI will stimulate creativity - but whose creativity? At stake in the issue, even before trademarks and intellectual property, even before the very notion of art, is the role society has invested the creative, in all its possible meanings. In our society, though often vituperated, plagiarized, underpaid, exploited and sidelined, creatives retain their social pedestal, that somewhat mystical aura that comes from the process, still all mental and human, of conceiving ideas and generating artistic content. Now those creative people are in danger of losing even that pedestal, substituted by software that not only replaces their work but devalues it, copying it for a tiny fraction of the cost. Which is a possible future drift of the contemporary drift that has seen, for example, the role of the creative director and the role of the designer capable of creating the actual clothes or even just the sketches diverge, according to an idea of mechanization of creative production (the chain of command of a large commercial atelier is a perfect example) whose final outcomes necessarily lead toward the complete automation promised by AI itself. The adoration we pay, for example, to Virgil Abloh for his innovative method of sending inspirations to his team via Whatsapp without drawing items himself (but few creative directors produce sketches and designs), breaking down and atomizing the various stages of creativity, is perhaps precisely what will pave the way for the arrival of fully digitized creative directions. The future, however, is unknown. We can, however, say with some certainty, quoting Victor Hugo: «Ceci tuera cela».