Since the beginning of film, two equally vital strands have animated it: the immersive fictional space constructed out of ‘reality’—as heralded by the Lumière Brothers’ The Arrival of a Train at La Ciotat Station (1896)—and the completely fantastical fictional space constructed with filmic sleight-of-hand—Georges Méliès’ The Trip to the Moon (1902) is perhaps its appropriate forebear. Across the history of fictional films, it’s perhaps the former stream (the Bazinian, after the great theorist of film, André Bazin) that has been most fertile. At least it’s the one I’ve most responded to, its ambition being nothing less than to reconstruct another total world from selected images of this partial one. It, for instance, gave rise to film noir, to Italian neorealism, to the New Wave, right now to the macabre realism of, say, Béla Tarr. The other strand, the one that cares less about physical space, about the ‘real world’, too, has developed into various forms of fantasy techniques, of VFX leading to the green screen and, now, virtual production.

The greatest scenes in film, ones that have a haunting power over me, have explored real space in some way or the other. Think of that last scene in Vampyr (1932) or even Psycho (1960), where Vera Miles’ character moves from the top of the killer’s house to the dreadful basement, or that slow, almost ritualistic walk around the room up to the portrait in Laura (1944). Entire films (say Barry Lyndon (1975) with its alternating claustral insides and desolate outsides of Kubrick or any film by Tarkovsky) have been structured around the making memorable, of different types of imagined space constructed out of real spaces. Perhaps the epitome of this might be Last Year at Marienbad (1961), an entirely spatio-temporal delight. Even the French New Wave would not have exploded if not for the basic respect for real locations and real spaces, morphed through trailblazing techniques, such as the jump cuts. The actual camera movement and actual location creates a sense of magic; I fear, however, a lot of the times that the second strand, when it ignores the real with an overdose of visual effects, can flatten out that true sorcery, and substitute for it a rather stale ‘irreal’ (which paradoxically in surplus becomes all too normal compared to the (still) strange enticements of the reconstructed ‘real’).

As an economist, I also wonder whether a select few VFX vendors will monopolise the virtual production market. But more importantly, I’m worried whether it’ll magnify a trend already dominant in cinema: the deflation of reality. That’s so because AI is evolving at a blinding speed. In 2018, my brother and I curated the world’s first AI art exhibition at a major art gallery, Nature Morte, in Delhi. At that time what existed was very different. We could have never predicted that AI would be where it is today. The ability of AI to mimic ‘natural’ parameters is only going to increase. (Now there’s a limit—the ability will not increase without bound—and that is data availability. At present, AI requires billions and billions of captured images, and we are close to the limit of image-making at which current models can operate). And the plague of fantasy images it unleashes can be so cheap as to make ‘real’ production relatively unattractive from a cost perspective. So cheap as to ultimately devalue reality. And not just with the big-budget films but also, and this is what I care about, the indies—films usually fortunately constrained by cost to explore real space!

Consider the two contrasting film industries in south India: the Malayalam and the Tamil. The latter (I’m speaking in broad terms, ignoring the wave of ‘Kerala’-type indies that are finding a growing niche in Tamil) might increasingly opt for virtual production because they churn out numerous ‘mass’ films; they’ve a different vibe, a different audience, different goals. But I also cherish the Kerala tradition—something like Kumbalangi Nights (2018)—which produces real and rooted films. And I truly hope the decreasing cost of virtual production will not devalue that ethos.

That said, though, it may not evolve that way! Digital photography, for example, didn’t replace painting. There’s something about actuality, actual spaces, the documentary medium, that’s become a fetish for some of us in our increasingly unreal world, with all these filtered images from Instagram, a wave of fake news, and what not. So a good portion of us will always value the kino-eye, the lived world filtered and remade by the lens, all the more. I do fear, however, that virtual production will shift the balance towards, what I find, very boring films or series such as The Mandalorian (2019).

I want to make clear that I am certainly not a Luddite, against technology in cinema in any way! Cinema and technical innovation go hand in hand. Some of the best films that I’ve ever seen were all black-and-white. But when colour came, everything changed. Of course, initially, many didn’t know how to handle it. But by the 1950s, say in Douglas Sirk’s melodramas, colour began to be used as an element of expression. So here is a technological development that really added to the aesthetic vocabulary of films in its Bazinian attempt to become a total medium mimicking all of reality.

So why is virtual production different? Well the two changes cannot be compared: colour brought us closer to the lived physical ‘meatspace’ reality: virtual production can very easily take us away from it. Virtual production can tempt the auteur to push the film into a fantastical, dream-like, ultimately solipsistic realm. It pushes us within ourselves, divorcing our points of contact with the outside world. Unlike the great tradition of filmmaking that has pushed us out into seeking fragile contact with what was once known as The Real World...

(Views expressed are personal)

Karthik Kalyanaraman is an economist who explores the intersection between AI and contemporary art

(This appeared in the print as 'Challenge Of The Virtual')