From Dust till Drone: Roomba Aesthetics and Non-Human Cinema
Neta Alexander / New York University
There is not much happening in this short YouTube clip: a glimpse of Mitt Romney talking on Fox News; a nameless white man sitting on the couch alongside his Labrador while absentmindedly staring at the television screen; a few wooden chairs and a kitchen table blocking the way. In short, a domestic scene that looks like an IKEA ad. And yet this five-minute clip, titled “iRobot Roomba GoPro,” has garnered hundreds of views since it was uploaded by a YouTube user named Grattan Heyward on January 10, 2011.
While Grattan’s contribution is an early example of the genre, by early 2017 more than 20,000 other “Roomba films” have been uploaded to YouTube – some of which reaped millions of views. How did this genre become successful? Why would anyone be bored enough to replace the kittens, panda bears, or adorable toddlers that occupy social media feeds with a clip made by a vacuum cleaner? While there are many possible answers to these questions, this short essay will contextualize Roomba aesthetic within the long tradition of non-human cinema. Moreover, I wish to argue that the surprising proliferation of this nascent genre cannot be detached from the growing scholarly and public interest in drones and “the rise of vertical perspective” as described by Hito Steyerl.  As above, so below.
While drones and automated vacuum cleaners entered our lives in the past two decades, the desire to free the camera from the human eye is as old as cinema itself. According to Dziga Vertov, the movie camera is not only an extension of our eyes, but a technological tool opening up a new kind of existential amplification: “Our eyes see very poorly and very little… the movie camera was invented to penetrate more deeply into the visible world.”  In Theory of Film, Zigfried Kracauer reminds us that humans were able to release themselves from the prison of their limited perspective and vision with the invention of aerial photography based on automatic cameras, which was first put to military use during the First World War. 
However, the current proliferation of “nonhuman cinema” – a wide-raging category consisting of GoPro films made by animals or objects, aerial footage shot by drones, or Sensory Ethnography Lab documentaries like Leviathan (2012) – is the result of two central developments in the Anthropocene: an enhanced awareness of the troubling aspects of human activity on both the environment and other organisms; and new wearable technologies that can easily turn an octopus, an eagle, or a vacuum cleaner into cinematographers and “content producers”. 
Nevertheless, there are substantial differences between Roomba clips and films made by animals. Take, for example, a short clip titled “Flying eagle point of view” which was uploaded to YouTube on 2014 and has since gained almost 10 million views.
At first view, our eye is attracted to the green landscape and the snowy French alps. However, the source of the attraction is not only the landscape, but also the documentation of the eagle’s head and feathers at the bottom right corner of the frame. The eagle thus simultaneously functions as the one who looks and an object to be-looked-at. This GoPro aesthetic could be found in numerous other clips shot by birds, mammals and even an “octographer” (an octopus-turned-cinematographer who starred in a 2015 Sony commercial for underwater cameras). The decision to include the head of the “cinematographer” in the frame is more than a technical constraint. It is an aesthetic and narrative device that serves as an indexical signifier. By showing us the cut off torso or head, we are invited to indulged in a magical world supposedly created without any human intervention.
In a similar manner, Roomba films challenge our anthropocentric perspective. In the “iRobot Roomba” clip, for example, a GoPro camera was attached to a Roomba vacuum cleaner. Instead of breathtaking landscapes, the mechanic gaze of the machine’s eye reveals the domestic sphere to be a frustrating maze. While it seems that this clip is devoid of any drama, there is a suspenseful encounter with a chair one and half minutes into the “narrative”. Our protagonist/filmmaker boldly approaches a set of chairs in order to dust the kitchen floor, but it soon realizes that the space between their legs is too narrow for its round, chubby, and mechanical body. A struggle ensues. In fact, it takes the Roomba almost 30 seconds – eternity in terms of YouTube clips – to release itself from the grip of the chair’s legs and roam forward with pride.
There is something weirdly reassuring in this short scene. It makes us more aware of the fact that the material world we inhabit is built around a set of assumptions, one of which is that human beings walk on two legs. The Roomba’s wretched journey through the house is a constant reminder that we are privileged in terms of our encounter with the spatial environment.
Moreover, the pleasure of watching the Roomba do its work cannot be detached from the knowledge that its human master sits on the couch while listening to Romney. The human is resting; the machine is doing the dirty work while struggling to find its way in an environment built and designed for humans. In this sense watching Roomba films resembles staring at what Anna McCarthy calls “industrial GIFs:” “Snips of people and machines at work, these hypnotic loops provide vicarious access to the linked sensations of precision, efficiency, and effortlessness associated with a job well done. We see drill bits go in, we see pencils sharpened. We see cakes on a conveyor belt sail through a waterfall of icing, again and again and again.”  While the little Roomba encounters obstacles and hazards, the viewer knows that it will eventually achieve its goal. After all, it was uniquely designed to clean our messy, cluttered households.
With their “hypnotic” quality, Roomba films can be seen as a mirror image of a more ubiquitous and prominent phenomenon: films, military footage, and short clips shot by drones. It is therefore noteworthy that the Roomba’s manufacture, iRobot, is also responsible for Packbot, a popular terrestrial military drone. In its effort to outsource various kinds of dirty jobs and clean the world from either domestic or human “dust” (e.g potential terrorists, refuges, and other unwanted entities), the company is currently developing “drone swarms” consisting of hundreds or thousands of tiny drones that can be used for surveillance, military strikes, or intimidation of civilian population.  As Malcolm MacPherson reminds us in his book Roberts Ridge, drones are capable of “giving commanders what they had only dreamed of: total situational awareness – making them like gods, omniscient and all-seeing”. 
Still, there is nothing “god-like” in a Roomba desperately fighting a kitchen chair. While drones produce a “vertical perspective” of the outdoors, what can be called “the Roombian gaze” is based on low-angle documentation of domestic settings. Drones see more than humans; Roombas see less. But there is also an uncanny similarity between these two technologies: they both produce a repetitive, disharmonious, and somewhat disturbing sound. Even when their function is different – saving time for people too busy to clean their home or playing an integral part in the War on Terror – their sonic presence is ominous: mechanical “white noise” of a machine that never sleeps, gets tired, or contemplate the moral or economic implications of its existence (do Roombas dream of electric sheep? I doubt it).
Drone footage often scares us because drones are invisible, omnipotent, and superior; Roomba films relax us and make us Laugh Out Loud because these machines – while capable of doing their work – are helpless and inferior. According to Steyerl, the ubiquity of technologies used for tracking, surveillance, and targeting has made us grow increasingly accustomed to what use to be called a God’s-eye view. The linear perspective that dominated the arts until the 20th century has gradually given way to the “vertical perspective” of Google Maps, drone footage, or satellite imagery. As a result, “perception is reorganized by warfare, advertisement, and the conveyor belt”. 
The fact that drones and vacuum cleaners are manufactured and developed by the same company can alert us to their oft-denied similarities. As different as they may seem, they adhere to the logic of “warfare and the conveyor belt” by sharing the idea that some jobs are too dirty to be left for human agents. Staring at a Roomba vs. kitchen chair showdown is far more entertaining than watching an undocumented immigrant cleaning a bathroom while fighting fatigue and disgust. These films dust away the earthly concerns produced by a system that relies on overworked and underpaid humans. Instead of looking fearfully at the skies, we can stare at the maze that we call home. Oh, bless the machines. They keep our houses, streets and hometowns “clean” and damn they are funny to watch.
Please feel free to comment.
- “In Free Fall: A Thought Experiment on Vertical Perspective”. In The Wretched of the Screen. E-flux Journal/ Sternberg Press, 2012:12-30. [↩]
- Kino-eye: The Writings of Dziga Vertov, University of California Press, 1984: 67. [↩]
- See Siegfried Kracauer, Theory of Film, Princeton University Press, 1960. Due to the limited scope of this essay, it will not explore the history of films made by automatic or algorithmic cameras. Notable examples include Michael Snow’s Back and Forth (1969) or Lars Von Trier’s The Boss of it All (2006). [↩]
- GoPro, the most popular brand of wearable digital cameras, was launched in 2002. However, the realization that these tiny cameras can be easily attached to either objects or animals had only gained popularity in the last several years. See Nick Paumgarten, “We Are a Camera: Experience and Memory in the Age of GoPro”. The New Yorker. September 22, 2014. [↩]
- “Visual Pleasure and GIFs,” in Compact Cinematics, Bloomsbury Publishing, 2017: 117. [↩]
- See John Teschner, “On drones,” The Iowa Review, Vol 43, No 1 (Spring 2013): 79. [↩]
- Qtd. in Teschner, 75. [↩]
- Steyerl, 2012: 22. [↩]