Following the last meeting, which was wasted arguing with a bridge, he convened Legislative & Rules Committee to order at 10:45 a.m. He was nervous about what the squirrels would have to say. It was an arduous and risky project, hallucinating the city back into existence from the sparse fragments of extracted sensor data. She never could have imagined that her old Bimmer could have committed to memory the late night drives and fast cruises around town, that training a car to dream was possible. The detective expertly dragged his gloved finger along the bookshelf and inspected the dust. The evidence was clear: they were going to have to start interrogating the Roomba. Tuning the birds fixed things for a while, but over time the ecosystem had become feral again. Their once melodic chirps and tweets, were now something more guttural and eerie. They were excited about touring the Unnatural Wonders preserve. After some orientation stuff – the tour guide had asked them not to use the term “obsolescence” – they set off to meet the machines. Finally, she thought, she could merge her love for gardening with her love for her city, and set to work pruning the decision trees. Reporting to the policing insects was a good job, but she was never fully at ease in their presence. “I’m doing swell!” He was shouting to the walls, trying not to look at them, waving casually to a person he pretended to know. Paying for coincidences was very addictive – and dangerous, if she was to suddenly run out of credit. Regardless, she stepped out into the traffic. You got used to avoiding the neighborhood’s thresholds; where one utopia rubbed up against another and the automation laws changed, you never knew how to properly behave.
This Town Has a Secret: Networked Colluding in the Internet of Things investigates secrecy as part of the fabric of a neighborhood, and how devices with artificial intelligence conspire. Drawing inspiration from mafia archetypes, the project takes the connectedness of IoT devices to an absurd future: a networked community of AI-agents who secretly control the neighborhood.
What could AI Noir as a genre look like? Using noir as inspiration and a neighborhood “mafia” as metaphor, the project manipulates the everyday narratives of AI in design and technology, as well as in neighborhoods and cities with this conceptual genre. The project imagines private investigators hired by the neighborhood council to solve a “mystery” involving the neighborhood’s networked colluding of IoT devices––a system of conspiring objects.
The project also uses AI Noir as a thought experiment for critiquing artificial intelligence on the neighborhood scale. Used as a conceptual strategy, as well as a visual technique, at the heart of this proposed genre is culpability. The final reveal in the design-performance is that the Boss is in fact a Roomba. If a cleaning bot were truly at fault though, how would events proceed? Would the device, company, or programmers be at fault? Recently, roomba maker iRobot made headlines for potentially selling spatial mapping data to smart home companies.
All in all, what are our limits of trust with these systems?
Training a Car to Dream comprises a series of machine learning apparatuses for training the neural nets of autonomous vehicles. The project imbues cars with the ability to dream on their own, reinterpreting what the term “dream car” is and means.
The project speculates new roles for how training a car could facilitate a shared hallucination with its occupants, one where the car fantasizes that it is driving very fast––a different mindset from its 30mph reality.
Can an autonomous vehicle be trained to dream––or more technically, to hallucinate and then simulate––a more thrilling, less uniform, transit experience for the benefit of its occupants? And what does it mean to be a bad driver in an autonomous vehicle?
A Committee of Infrastructure interrogates the issue of agency and representation within the domain of machine learning and artificial intelligence.
The imagined council meeting takes place in Los Angeles wherein human and artificial representatives voice conflicting positions, ideologies, and motivations. What are the interactions between AI systems and people at the community and local government level, and which previously unrecognized communities are now given agency through AI representatives? In Sweden, engineers have created a network of Stockholm’s aging bridges, some with up to 72 different sensors. This creates a massive stream of data and tweets about the city’s infrastructure wear and tear. How can this data become part of the civic dialogue?
Using the familiar forum of a city council meeting, the project considers how humans and AI entities interact and negotiate with each other in a local government setting. Will our AI civic participants be just as narrow and petty as humans, or is another outcome possible? The assemblage of biological and artificial voices inhabiting the civic space reexamines the idea of personhood and through this reexamination, all in turn are complicit in the political.
AI Upkeep proposes that the hidden intelligent systems controlling neighborhoods are made open and legible to the public in the form of physical, manipulatable tree-like structures. How might pruning these civic interfaces literally and figuratively reshape the actual neighborhood?
The new typology of public space proposed in the project combines the mechanical qualities of urban dashboards with the permeable and spatial qualities of public parks. The new “parks” contain “AI topiaries” that physicalize what is otherwise invisible to citizens: the algorithms behind the town’s automated intelligent infrastructure – the decision trees and neural nets that are involved in the public decision-making and sense-making, that literally and figuratively shape an AI-embedded city. By externalizing civic AI systems into public space, AI Upkeep aims to create active and participatory modes of collecting data used to train these systems, as opposed to passive accumulation and infinite siloing of data.
How does one organically update their city? The project takes inspiration from the popular TV show Mr. Rogers’ Neighborhood and asks how we can do what Mr. Rogers did for television and demystify the medium of artificial intelligence? When the software dimension of a city is controlled by civic AI, how can we counteract the opacity of civic AI and ML processes, in order to make those systems more transparent and even participatory?
Rules of Utopia imagines multiple Homeowner’s Associations (HOA), each with different rules governing the behavior of intelligent devices within the town of Utopia. What are the conflicts that could arise between districts with different sets of rules, and how do the HOA thresholds of autonomous device regulations respond accordingly?
Envisioning a district cheekily named Utopia, the framework of a Homeowner’s Association helps expose the challenges of hyper-local AI, rule making, and human-computer integration at the neighborhood level. What are the rules that AIs would set up regarding the maintenance of common areas and voting processes? The project speculates on the increasing role of AI systems for safety networks and the types of utopias the companies behind them are desiring to create.
Intelligent Devices Retirement Preserve imagines a parkland where intelligent agricultural machinery can continue to roam and interact with people after decommissioning. The project considers roles for specific classes of smart devices beyond the end of their designed obsolescence, particularly autonomous farming equipment which will have acquired a unique pastoral data set through a life of tending crops and livestock.
Drawing inspiration from junk drawers filled with old phones and broken drones, if a place were to exist for decommissioned devices, where would they go and what would the relocation process look like? What are the new roles of devices given that they have unlimited “free time” because they are no longer with their previous human owners? Elements of a device’s past history could now become a hobby, such as playing minesweeper as a way to pass the time.
The retirement preserve proposes a more sentimental approach for how we deal with end of life care for our most beloved, warmed-over machines.
When AIs Go Feral imagines how the life of a neighborhood is modified by artificial autonomous agents gone wild. Inspired by a real life story of a flamboyant real estate developer who imported non-native species of birds onto his private ranch, the project uses the suburban birdsong soundscape to explore a constantly mutating and evolving hybrid ecosystem.
What are the consequences of AIs that become feral in a neighborhood? Would AI entities alter the sounds of their communications in reaction to the natural environment? This reimagined history explores what it means to learn to live with alterations to the landscape and how even our ears will eventually adjust to the sounds of our new, noisy ai-neighbors. Like a game of Telephone, these AI birds audibly communicate with one another, however, their messages become distorted over time, oscillating between what sounds like digital sine tones and bird-like calls. At what point are the synthetic sounds perceived as natural? What kind of influence will an environment’s soundscape have on AIs that inhabit the space? What are the broader implications of, uncontrolled, “non-native” AI species becoming part of the neighborhood?
Listening City explores human relationships in the imagined context where infrastructure can, literally, hear what you say and “usefully” intervene by reacting with spawns of emojis indicating the person’s positivity and negativity analysis. Inspired by developments in sentiment analysis, the project looks at how radical sensitivity can be embodied by AI city infrastructures and how this extreme “smartness” can change the behavior of even our most passing comments.
Sensing classification systems and the urge for optimizing everything can be problematic and are mostly inherited from non-user focused data scientists and software development teams. If the city’s walls are always listening and classifying our conversations, how often do they get it right? How can the AI-powered walls and people differentiate between a genuine conversation and one that is not? Do public interactions become performative?
Insectile Indices considers how electronically augmented insects could be trainable to act as sophisticated sensors, working in groups, as part of a neighborhood policing initiative.
In 2007, the Defence Advanced Research Projects Agency (DARPA) asked America’s scientists to submit proposals to develop technology to create insect-cyborgs. Building off this frightening initiative, this project inverts the sinister military connotations of this proposed future and imagines instead an aesthetically pleasing utopia where sensored insects work towards the public good of humanity. Insectile Indices also plays with the idea of aesthetics in our techno-futures: if these sensored insects are approachable and pleasing to the eye, are we more apt to silently turn the cheek to more pervasive surveillance?
The project is partly an investigation into the ethics of this controversial idea, but also an aesthetic exploration of such a deliberate alteration to a wildlife ecosystem.
The Hallucinating City imagines a city rebuilt, conjured back into existence by strategically “hallucinating” forms from fragments of excavated media and metadata.
The project is loosely based off the author’s own story where she was once familiar with a building in the Czech city of Brno, but over time her familiarity with the structure began to fade. The recollected images seem absurd and abstract when isolated, but the series of the author’s collected mental pictures from the fragments of the edifice are used to output a hallucination “completing” the building in its entirety.
Although the machine learning tools investigated in this project are used for a speculative archaeological reconstruction and utilize the aesthetics of hypergan image training, how can we fully complete images of what was lost with different variations of remembrance? The project explores the blurry line between nostalgia and AI hallucination, and the powerful, yet contentious role machine learning can play in materializing something tangible and concrete from a transient and fragmented past.
Optimization for Mundanity imagines civic life where extreme optimization extends into the fabric of the everyday mundane. What absurdities come about in a system where every aspect of life can be optimized, including your moment by moment location on the metro?
The project speculates that an AI-powered public transportation system can get you the absolute optimal location during your commute. The commuter uses on-demand micro-payments to access the app’s crowd prediction algorithm and reserve the best spot in terms of comfort and access to the exit. Going further, the system allows additional “bribes” to upgrade the commuter’s position to an even better location or seat, while bumping another passenger.
Optimization for Mundanity pokes fun at the incessant progress of optimization and solutionism as manifested through mobile applications. It asks if the saving of extremely minor levels of comfort and time are worth the mental, technological, cultural and monetary costs.