What will be the first sights of the conscious machine?
How will it apprehend the world around us?
What will it see?
Considering the disastrous path humanity has chosen to follow over the last hundreds of years, we could well imagine that the birth of a fully conscious AI being will be marked by images of vast destruction, hurricanes, floods and other environmental catastrophes.
All over the world, there is now well over 250 millions cameras, webcams and surveillance cameras, recording our reality 24/7. The sheer quantity of data made available from hundreds of millions of locations can only be viewed or at least processed by machines, and possibly one day only fully integrated in a kind of super vision by a yet to be AI super-consciousness. An hyper being with 250 million eyes. The gaze of these cameras has become the prism by which we are able to theorise and disengage from events that otherwise would harm us. Unharmed, unaffected und undisturbed the camera keeps recording.
But how do they see?
This is what we set out to understand; to bring computer vision back to painting, to try to internalise the machine way of seeing, its brute logic of dissecting the field of view in millions of tiny pixels, and the range of colours it perceives. The paintings are deconstructed into pixels of 16 shades of Red, Green and Blue, reconstituting the way that networked cameras, CCTV, aerial drones and all digital cameras see and record the world: in raw format, in digital RGB. They depict publicly accessible webcam imagery of extreme natural weather conditions and events as a technique for visualising the world with the detached, indifferent and disengaged eye of CCTV or 'webcam vision'. In the process, the act of painting become nearly entirely logical, by numbers. A tedious, repetitive task, the mixing of the colours a scientific endeavour, the application of the paint requiring a mechanical, robot-like precision. And yet, as one applies the colours, pixel by pixel, moments of colour harmony emerge, instants when the subtle, unforeseen combination of greens, red and blues arise, and all the while one is transported in a curious, ambiguous state of mind. Reduced to the original definition of the computer, in a state of complete logical concentration, the act of painting, through its repetitiveness perhaps, brings you into a meditative, transcendental space. You become the machine, and as the picture emerges, a visceral understanding of what a conscious machine could feel deepens...
The arrangement of the coloured pixels is very specific to networked cameras and called the Bayer filter array. On closer look the pixelation is made up of a very specific grid of red, green and blue that are created by a special pattern of red green and blue filters that are overlaid over the camera's sensors to cover each of the pixels.
Once a camera has recorded an image in this way, and to then obtain a usable full colour image that mimics what the human eye sees, algorithms in the camera or in the computer interpolate the red, green and blue values for each pixel. The strangely green hue of the paintings is the result of there are being twice as many green pixels compared to red or blue, a method to counter the greater sensitivity to green light of the human eye.