Google Dubtechno Now

Vision, Attention, Consciousness

27th of September, 2021

Some artificial intelligence researchers believe that building a machine that has consciousness is relatively simple, but currently beyond our reach because we have yet to create a sophisticated enough system of perception. This sounds implausible at first but when disambiguating how the word “consciousness” is used it becomes much more compelling. There is a consciousness that entails a subjective experience of qualia, and there is a consciousness which is more broadly just a description of reflexive awareness—aware of being aware. A system exhibiting the latter sort of consciousness seems both obviously not that difficult to build but also like cheating, or missing the point. In 2015, when researchers at the Rensselaer Polytechnic Institute built a robot that could solve the wise-man problem, no one considered the hard problem of consciousness solved. But, it seems reasonable to expect that the difference between something with an internal, subjective experience of qualia and something without this experience hasn’t got anything much to do with the awareness of oneself. It seems to hinge on having a rich enough system of perception, and a powerful enough system of processing all that sense data in a multi-media language.

There's an image stuck in my mind, I don't remember what film it's from: some hands fiddle with a deck of cards, behind the hands is a table with some bits of tobacco on, and as the table stretches further away into unfocussing, a VHS tape helps delineate the exact extent that firm outlines become impossible at that distance from the lens. Even without the shuffling of the hands and the cards, which I assume is the intended source of dynamism in the shot, the operatic drama of the scattering of light that pushes these surfaces out of focus is endlessly compelling. The diagonal shift creates zones of noise and blurred colour which flatten out and sharpen into cones. They ripple as shadows passing over the surfaces disturb the grain of the flat image we are seeing. Because, we must remember, the image is flat. We are used to thinking of imagery and photography the way we think of sight. We see a landscape photograph and place ourselves in it, or in front of it. We disambiguate between figure and ground. But the layered transient blurs and fuzzes, the cones of light that suggest shadow and distance and solidity, are bounded on all sides by an oblong that reminds us that the operation of the light within the frame is a story.

Seen with the eye, the drama of the dissolving VHS tape would be hidden from us by psychological necessity. In his essay film In Search of a Flat Earth, Dan Olson is in his studio looking at high quality footage he has taken from a lake that demonstrates the curvature of the Earth. He admits that the scene moves him.

“You can see this just with the naked eye when you’re out there on the beach. You can pick a thing on the far shore and just kind of look at it and crouch down, get really close to the water’s surface and see it get occluded. But… but seeing at this magnitude… it feels profound.”

Such is the hyperreality of film, or of documentation in general. Olson spent hours out at the lake using his own eyes to inform where to put his camera, but it is when he gets home and looks at his own footage on a giant monitor that the full weight of the imagery begins to affect him. The act of looking at anything at all necessarily produces these incredible optical effects—the shifting dance of parallax and blurring and bleeding star shapes of lights seen through moisture… these things happen around us all the time but it seems to take the impact of freezing them for us to realise the profundity of experience. In the moment we tend to spend so much time processing, disambiguating the image into coherent and discrete things like surfaces and actors, that we rarely have time to appreciate the staggering complexity of the multi-media language our brains are able to encode these representations with.

Last week I was watching an archived live broadcast: a camera set up to focus on two specific people sat stationary as they stood up and walked out of the frame. Behind them was an image of a convention centre, and fifty people wandered through the scene, possessed by their own motives for doing so. Like Olson, something I took for granted as part of my experience was shown to me as an image, and I was moved. My God, every single one of those people, as they moved through that space, contained an internal universe of experience, and I’m watching them, and they are contained in me as I do so.

This way of thinking privileges perception immensely. Sight is just one facet of this—of course blind people are still conscious—it is the internal processing of experience that produces the compelling narrative that constitutes our inner experience. Once a system can use its reflexive awareness to manage its attention, it has this giant mountain of sense data to sort through. It has to produce a system of attention—and as soon as it is aware that it is using a system of attention to choose what to process, it is thinking consciously. It produces narratives about the world and tells itself stories about why certain things are important. These stories might only be as simple as a model of its goals, but another way of describing that simple act is the creation of representations of reality in its own mental language, which is exactly what qualia are.

The world does not behave the way we perceive it. The blurring of light that suggests the relative distance of surfaces is something that we have to form internally as a model, the same way we have to interpret the jostling confluence of water molecules as something as coherent as a wave. Trying to think about the way the world is, the incomprehensible operation of physics reveals the extent that we live in a dream world. Nothing we perceive or think about resembles anything that is happening physically. This for me makes the mechanical diffusion of light, and arrangements of surfaces, something uniquely human, because it doesn’t exist outside of my experience of it. The image of surfaces veering away into distance and spatial ambiguity and abstraction is at least as human as an actor’s hands pretending to shuffle cards.

It might occasionally take a recording of a full convention centre to bring me close to tears, but images of people are accessible to us whenever we use our awareness to organise our attention. Dan Olson could choose to be moved by the image of the tree’s occlusion by the Earth’s curvature, by forgetting for a moment that he is a filmmaker trying to capture the image and simply taking a few seconds to really look. I can choose to be moved by people that move around and past me whenever I focus on the devastating fact that I am surrounded by heads full of consciousness, organising imagery.