In March 2018, on a cool Sunday night in Tempe, Arizona, a woman was struck by a Volvo SUV while trying to cross the street with her bicycle. She later died of her injuries. The police were perplexed as to how they would proceed with a potential criminal investigation. No one was driving the car.
The vehicle was a self-driving prototype on a test route for Uber, monitored by a ride-along operator who had no time to intervene before collision. The victim, Elaine Herzberg, will never know that she was hit by a car driving itself. Her death was classified as the first self-driving car pedestrian fatality (an operator was killed in a crash due to a glitch in 2016). For Herzberg, it seems that she was simply in the wrong place at the wrong time. Journalists morbidly mourned her while simultaneously utilizing the death as another chance to rhetorically muse about the dangers of automation and Silicon Valley ubiquity. Needless to say, no complaint was filed, as Uber was all the more ready to settle out of court.
Shortly after the tragedy made the media rounds, The Guardian reported that Arizona governor Doug Ducey was not only aware of Uber’s surreptitious self-driving tests, but actively enabled them in secret. The state of Arizona intentionally kept its citizens unaware of the fact that these vehicles were driving their streets at night. Behind closed doors, Uber had promised Ducey money, jobs, and technological innovation for his state in exchange for a hushed series of tests. The whole gambit played out in almost comically corrupt fashion, involving a hamstrung oversight committee and an attempted gubernatorial pay-off with an Uber t-shirt. Not surprisingly, upon this information becoming public, a jilted Ducey unceremoniously revoked Uber’s right to continue their autonomous vehicle research in Arizona, and other self-driving tests around the United States came to a temporary halt.
Ducey is not the first politician to become embroiled in a Faustian deal with Uber. In fact, lean platforms such as Uber have perfected the art of subtle exploitation. As Nick Srnicek describes in Platform Capitalism, “lean platforms operate through a hyper-outsourced model, whereby workers are outsourced, fixed capital is outsourced, maintenance costs are outsourced, and training is outsourced.” Such deregulated models are appealing to government and big business alike. Freed from traditional assets, these platforms are hyper-cost-efficient, and the promise of economic stimulation via “the sharing economy” has become a guiding theme for Uber, Lyft, Airbnb, and others. Workers are utilized on the basis of immediate need, and companies that contract them no longer need to provide the usual support systems of a traditional employer. Returns fund research like the tests carried out in Tempe, which seek to develop AI replacements for their current human counterparts.
Death and destruction at the will of malfunctioning autonomous vehicles certainly feels like a futuristic nightmare. However, it is not difficult to imagine an intelligent machine going against our demands and surpassing our control. We are conditioned to be trepidatious of technology by science fiction, political systems, uncanny valleys, and economic anxieties. Within this dynamic, New York-based artist and programmer Aashish Gadani’s 2018 video, DARPA Grand Challenge, stakes its ground.
In the video’s opening scene, two men are seen fishing from a pier. Suddenly, like the jarring reveal of a movie monster, a red sports car flies into frame, scattering seagulls and squealing tires. Smashing into various pier detritus, the onlookers groan and shake their heads. In a clutter of scenes, each only a few seconds long, the car appears to be simultaneously wreaking havoc at a harbor and on a mountain road. The scenes become enmeshed, and it is difficult to find sure footing. Once things seem to be approaching cohesion, there is another jumpcut, and the red sports car could be anywhere.
A primary facet of Gadani’s visual work is the conceptual reframing of found footage and other media. These reframings often foreground the troubles of technological accelerationism, with a focus on surveillance and conspiracy. In 2016’s Faces of Boiler Room, Gadani utilized the software OpenCV to perform facial recognition, culling through YouTube videos of live DJ sets uploaded by the music platform Boiler Room. The result was a website that accumulated 75,000 facial profiles into a massive digital collage. Upon initial interaction, the chaotic sea of pixels seems unreadable, but zooming in, one can see faces gleaned from multiple audiences (and the occasional misidentified water bottle) gathered uniformly in their cells.
Whereas Faces of Boiler Room is structured around static, monolithic mosaic, DARPA Grand Challenge shifts gears into filmic montage, utilizing repurposed footage solely from The Fast and the Furious film series. The artist describes the plot of the almost hour-long video as “a future in which humans are in conflict with sentient autonomous vehicles. Presented in 3 acts, it follows the development of these cars, their progressive disregard for collateral damage and finally…fully actualized war against humans and their car allies.” Gadani collapses the linearity of the source films’ already convoluted story arcs, compiling clips into a spasmodic series of the chase-crash-continue scenes for which they are infamous. From the rubble, Gadani loosely formulates a narrative that pits familiar characters, now rendered obsolete and contextless, against speeding sports cars. The lynchpin of the project is the depersonalization of these vehicles: when the driver is unseen on the screen, Gadani recasts the cars as sentient machines seeking revenge.
DARPA not only locates its lineage in early avant-garde collage films, but also finds company in the realm of YouTube compilation videos. A quick YouTube search for “car chase scene compilation” produces 1,710,000 results. The memetic quality of The Fast and the Furious franchise, ostensibly some of the most recognizable action films of the early 2000s to present day, makes the video all the more prescient. Swaying between modes, the work is indicative of both death drive and death dérive.
We see cars racing and swerving in a parking garage ouroboros, punctuated with bad CGI montages of exploding combustion engines and the twisted metal of flipping cars double-helixing through Tokyo traffic. The insectoid drone of engines are a constant noise interrupted by fragments of tracks from Ludacris and Kid Rock. As the video progresses, Gadani gradually introduces clips of technological GPS and tracking interfaces, implying the visualization of the sentient car consciousness. “When we find out what it is, we’ll know what we are up against,” Vin Diesel mutters.
The vehicles annihilate the infrastructure that contains them to lanes, parking spaces, and freeway ramps. In a sense, these vehicles were once part of this infrastructure, but have now turned destructive, and are looking to maximize damage to the architecture of their containment. In one of the most compelling scenes, several cars (including a glaring Jeep Grand Cherokee product placement) burst forth from the glass storefront of a dealership to join in on a catastrophic metropolitan police chase. As citizens watch on in horror, object destroys object. In the madness, sentient vehicles and non-sentient vehicles driven by people become indistinguishable. They crash all the same.
Gadani’s work is emblematic of a particularly American dromology. Defined by Paul Virilio as “the study and analysis of the increasing speed of transport and communications on the development of land-use,” dromology makes the dynamic cityscape its major site of inquiry, and the forms of acceleration that simultaneously contain and hyper-mobilize citizens. Benjamin Bratton, in his introduction to Virilio’s Speed and Politics, summates: “Dromology (from Gr. dromos: race course) is this government of differential motility, of harnessing and mobilizing, incarcerating and accelerating things and people.” For Virilio, such technologies are typically extensions of governmental control, but in the context of giant American tech companies like Uber, the companies that own these technologies exceed Virilio’s expectations. In DARPA, the prosthetic technology of self-driving cars wrenches itself from both its systems of control but also the masses, forming a humanless revolutionary armada that turns all of the potentials for dromological incipience into a mechanized self-destruction.
The video not only reflects dromology in its content—a speculative crisis inflamed by rapidly evolving technologies—but also the form in which it is presented: one long, unrelenting action scene that pummels the viewer with Hollywood blockbuster tropes. In this concentrated dosage, the mechanisms of these films become most apparent. The Fast and the Furious films are lucrative box office productions upholding the fetishization of luxury automobiles as a signifier of class, ‘coolness’, and hypermasculinity. The perverse desire of glamorized expansion (road movies, Route 66, the “need for speed”) is certainly a baseline reasoning for the films’ popularity. What other desires are viewers of The Fast and the Furious projecting onto these autonomous vehicles? In a sense, this thirst for destruction is a complicit thirst, unfettered by culpability and real consequence. Popular cinema thrills by offering the impossible and unobtainable. Gadani knots all of this up, sideswiping the symbolic in favor of disorienting, depersonalized accumulation. He highlights the absurd excess of violence that automotive culture perpetuates through an enfilade of pyrotechnics and stunts. The artist, aware of our anxiety of losing control to our own machines, creates a possible world where cars finally outwit us with computer intelligence. In DARPA, human control can never be guaranteed.
Gadani’s video crafts an adrenaline-pumping yet ironically meandering dystopian future. At a certain point, one’s threshold for jam-packed pandemonium is reached. Excitement is completely drained from the spectacle. Raging car chases become as docile as slow television. Any plot twists or climaxes become unknotted, flattened. Such is the potential hazard of overexposure. Whether or not this was Gadani’s intended experience, I find it to be an astute, if not disturbing, condition.
As transportation technologies continue to develop, it seems that most of us are simply along for the ride. In the post-Herzberg era, the streets have a new producer of fear: unrecognizable at first glance, they blend into traffic, scanning and gathering spatial data in silence. Summoned by an app on your phone, one arrives empty and welcoming. All the way to your destination, you daydream: the car door locks, the engine revs, and with a methodical turn of the wheel, you begin to drift.