There Is a There There
"Every photograph is a certificate of presence."
--Roland Barthes, Camera Lucida
K-Pop performer Bada in Yongsan in 2022. It is difficult to tell where Bada’s hair and the enhanced version ends and begins, or where her makeup was and the digital stuff begins, but we can still tell Bada was there.
It’s Not “AI”
Many hands are wringing over “A.I.” But the problem is that actual artificial intelligence isn’t actually here yet. The term “AI” seems to be invoked and employed to describe whatever wondrous automated processes come into play to do things that would normally take a human a lot more time to carry out, such as when Chat GPT talks with us and give intelligible responses — often responses that seem to do work for us, such as “write me an essay about Dostoevsky’s thoughts on the human condition”, and when we see the output, it we love it because it passes muster. It’s the output we wanted, so it must be self-aware, right? It must have some thoughts about Dostoevsky, right? It knows about him, right? It told me something!
But that’s wrong. It’s just the best response the complex algorithm could give us. It passes the Turing test, which tells us that “if a machine can engage in a conversation with a human without being detected as a machine, it has demonstrated human intelligence.” (note)
The thing is that none of this is a function of a self-aware intelligence that sits and understands things. That’s why AI doesn’t get hands. It doesn’t know what they are; or what they do. So it doesn’t know how to draw them consistently — yet.
In the same way, “AI” didn’t make deep, creative choices in doing this makeup. The algorithm in FaceApp has obviously been well-trained on lots of examples. It knows how thw colors on eyes are drawn, and how the black lines around it are supposed to look, based on thousands of cases it has seen. And it has obviously been trained to know that it can’t be drawn over hair, because that is against the rules of its training. So it knows how to deal with hair. And it takes its cues from the existing makeup on the face it is dealing with, so it’s helpful to the program, as well as for matchingup with reality, which helps give an answer the user finds desireable, just like Chat GPT does when it writes your class term paper for you. It’s what you want. Ad most of the time, it does a pretty good job, until you realize it has just started making things up to finish out the response the user wants. Facts (or reality) don’t matter — just finishing the output. Which is why you get essays with made-up, non-facts sometimes, or garish mascara drawn over the hair if the model is at a strange angle. But if the model’s hair isn’t too much in the way and she’s looking straight at the camera, the “AI” makeup does a spectacular job. Especially if you don’t ask it to do too much.
But the thing Barthes was saying about a picture was that there was a there there at one time, a reality to which a picture refers. That is how a picture has always been different from a painting, or another piece of art generated completely from out of the blue, out of someone’s mind. A picture is “a certifcate of presence” in that something happened, that person was there. It is evidentiary; it is proof. Sure, it can be altered, it can be digitally adjusted to look a bit different, and the small details can be changed. But the reason I don't have to worry about the models, hands looking funny or backwards or nonhuman in the picture above is because she was actually there, and the model actually occupied that space in the way that she did.
Altering the Aesthetics of Reality
Nowadays, we live deeply steeped in hyperreality. In short, we live in a world in which we can’t tell the original things from the things (signs and symbols and other things) that represent those original things. And on top of that, once we’re so used to the sign/symbol that represents the original, or don’t care when a copy of a copy of a copy is indeed a copy, then the original loses meaning. Such a situation used to be so confusing that it could cause fainting spells, but nowadays, since we live inside this situation, it’s quite easy to imagine. On the daily, we interact with one another with digital duplicate of ourselves, digital avatars that we take to be “me” and which you take to be your friend “Jiwon” or “Frank.” it may be their profile picture+their name+ID, but you take it for your friend and interact with them with your own avatar. We base a lot of social interactions in certain digital spaces such as Instagram or Facebook, and we often don’t see that person for days, weeks, months, or even years. Jean Baudrillard talked about this slippage in Simulation and SImulacra,
In the realm of hyperreality, where the line between what is real and what is representation becomes increasingly blurred, “artificial intelligence” plays a pivotal role. It contributes to the creation of images and avatars that serve as (partial) stand-ins for reality, blurring the boundaries between the original and its representation. The algorithmic processes behind technologies like FaceApp, while not possessing genuine intelligence or awareness, are adept at mimicking reality based on vast datasets and learned patterns. These technologies create convincing alterations to images, such as makeup enhancements or facial transformations, leading us to question the authenticity of what we see.
In the ever-evolving interplay between artificial intelligence and our perceptible reality, it becomes essential to grasp that even amid the transformative prowess of AI enhancements, a foundational reality persists. As Jean Baudrillard astutely noted in his reflections on Simulation and Simulacra, "It is the real, and not the map, whose vestiges persist here and there in the deserts that are no longer those of the Empire, but ours."
The seemingly hyperreal representations crafted by AI algorithms are intricately woven from the fabric of genuine human experiences, encapsulated within datasets that serve as the crucible for these intelligent systems. In marveling at the digital avatars and augmented images, we must not overlook that their essence derives from the tangible reality we collectively share. In this symbiotic relationship between AI and the base reality it emulates, we uncover a profound connection that echoes Baudrillard's insights — the vestiges of the real persist, asserting their presence in the intricate dance between the simulated and the authentic.
The vestiges of the real persist. Bada (@d.a.d.bada) and I were real people in a real abandoned (and condemned), old house, and this was a space that conveyed a real mood, which informed the way she comported herself and posed her body. Her makeup was real, as was her hairstyle, which, despite being enhanced by AI makeup and hair, gelled into a real, organically whole moment.
Baudrillard argued that in a hyperreal world, the simulation overtakes reality, and the distinction between the two becomes irrelevant. The digital avatars we engage with on social media, the manipulated images produced by AI algorithms, and the augmented realities we experience through various filters all contribute to this complex web of hyperreality. As a consequence, the original loses its significance, and the representation becomes the dominant reality.
The impact of AI on our perception of reality extends beyond image manipulation. ChatGPT, for example, mimics human conversation skillfully, providing responses that may seem genuinely thoughtful and insightful. However, it lacks true understanding or self-awareness. It excels at generating coherent and contextually appropriate responses based on patterns learned from extensive training data. Yet, the absence of genuine comprehension raises questions about the nature of our interactions with these intelligent systems.
As we navigate the hyperreal landscape, where the boundaries between reality and representation are continually blurred, it is crucial to recognize the role of artificial intelligence in shaping our experiences. The "there there" that Roland Barthes spoke of in photography is complexified by the hyperreal simulacra crafted by AI. While new technologies offer myriad, new capabilities, it is essential to maintain awareness of their limitations. As we hurtle headlong into the age of AI, the fact of the "there" right there in our lived experiences. Yes, wholesale simulations and a clusters of simulacra often complexify the question of where reality ends and augmentation begins, but the power of the picture — a real picture and not an all-out AI simulation — lies in it being tethered to reality, a “certificate of presence.”