GIFT A MEMBERSHIP
A year of exhibitions, theater shows and concerts to enjoy with your favorite people. Join today or gift a membership.
Triennale Milano
© Kim Albrecht

Artificial Senses told by aesthetic researcher Kim Albrecht

February 4 2022
The machines so close to us
Contemporary culture is unimaginable without the machines that surround and guide us throughout the day. Google search results frame our knowledge, the mixes Spotify creates for us shapes our music tastes, and Amazon recommendations influence our consumption. All these systems present themselves as unknowns to us through their slick interfaces, obscuring the actuality of these systems. 
This strange new world became an intrinsic part of our reality in a brief timeframe. Companies that are part of this economy developed strategies that turn these new technologies into something that feels natural to us: 'It just works,' as Steve Jobs liked to say. (Distelmeyer 2017; Moisescot 2009) Interface design is the discipline that turns the machine into something that creates this natural feeling. But if we want to live with these devices and understand them, we cannot merely rely on the machines becoming something easily understandable to us. We need to develop an understanding of how these devices are different than us, as humans, and at the same time understand how technology is always culturally created by us.
With machine learning and artificial intelligence being such buzzwords nowadays, not having an understanding of how these machines experience the world is dangerous. How can we live and interact with these alien apparatuses that we set into the world, if the only thing we know about these machines are constructed interfaces that make the machine feel so close to the world we already know?
Computer-Human Interfaces usually try to be as human as possible. Conversely, in this series of works, the interfaces seek to be as close to the machine view as possible, creating images that are strange and difficult to interpret for the human viewer, but entirely natural for the machine.
© Kim Albrecht
From Operational Images to Images of Operation
In the early 2000s, documentary filmmaker and artist Harun Farocki distinguished a new kind of image. He called them 'Operational Images'-pictures that are part of a technical operation, a process in which they are needed to for fill a specific goal. (Distelmeyer 2017; Paglen 2014) In the movie 'Eye/Machine III' (Farocki 2003) Farocki showcases how the machine sees, how it experiences the world, and how it uses operational images to do so. For example, in the third part Farocki goes through the process how a laser-sensor examines a bridge; how that bridge is read into the computer as a three-dimensional model. The next cut shows different machine perspectives: a photograph, a 3D model, a satellite image overlaid with graphics, and airplane video footage including markings by the machine. As Trevor Paglen points out (Paglen 2014), the problem is the machine does not see-humans see. Our machinery only calculates. They do not need operational images, as they do not rely on visuals to sense the world. Images showcasing how the machine sees always rely on an interface, a mapping from the machine to the human world, from understanding the world through calculation to understanding the world through human senses.

© Kim Albrecht
© Kim Albrecht
© Kim Albrecht
As humans, we are very good at seeing. Our brain heavily relies on the visual. The machine, on the other hand, is based on binary numbers. Computers experience the world through mathematics. In this sense, operational images created by machines for machines are invisible. The visual mapping, the lines, and boxes overlaying surveillance cameras to showcase face detection are visuals that bring the human back in the loop but are not needed by the machines. They are not the images of operation, as the operations are only strings of numbers. The machine does not need images; images are just another stream of data. The translation-the mapping-to the visual is only necessary as an interface to us humans.
Data visualization is a negotiation ground of the two styles of experience between humans and machines. While visualizations are based on numbers and mathematics, they transition into forms that humans can experience with our most efficient sense, our eyes. Visualizations are operational images themselves-images in operation, with a cause, with a goal, images used within an operation. We do not open our map application on our phone to enjoy the beauty of the map, but rather to get from place A to place B, to fulfill an operation. We do not draw a scatterplot for the beauty of the plot but to find outliers, see trends, and to make sense of the data. There are exceptions but to a large extent visualizations help to explain or explore data and make it humanly readable, making it usable within an operation between machines and humans. The images created throughout this project are explicitly not doing this. That is, they are not visualizations for a utilitarian operation. What they do is map the operation of the machines and make it visible, in a manner that is as close as possible to the machine experience.
© Kim Albrecht
© Kim Albrecht
© Kim Albrecht
The here presented graphics are not 'Operational Images' but rather 'Images of Operation.' They make the machine operation visible, turning it into something that we can experience, but not necessarily act upon it. Such a counter logic of what visualization can be and could do opens up new paths for the field. Rather than using computation, data, and visualization to look into the world, to understand nature and culture, we use visualization to look into the machine itself, into its mechanics and infrastructures that allow for its creation in the first place. Images of operation make the machine visible in a way that is close to its mechanics. That is, they create visualizations that are not humanized, not designed to be as humanly accessible as possible, but as close to the original input as possible.
Credits
Bibliography Distelmeyer, J., 2017. Machtzeichen. Farocki, H., 2003. Eye/Machine III, Auge/Maschine III. , 63 min. Farocki, H., 2012. Parallel I. Available at: https://www.youtube.com/watch?v=Yzc3OPc8gUM. Moisescot, R., 2009. It just works. Seamlessly. Available at: https://www.youtube.com/watch?v=qmPq00jelpc. Paglen, T., 2014. Operational Images. , (59). Available at: http://worker01.e-flux.com/pdf/article_8990555.pdf.