Insectile Indices, 2017


Featured at
Ars Electronica 2017
Post City, Linz, AT

AAAI Conference, 2017
Standford University, CA, US.

Insectile Indices is a speculative design project that considers how electronically augmented insects could be trained to act as sophisticated data sensors, working in groups, as part of a predictive policing initiatives in the city of Los Angeles, 2027. This project is not only an investigation into the ethics of this controversial idea, but an aesthetic exploration into the deliberate alteration to a natural wildlife ecosystem of insects and the potential reshaping of an urban landscape.

In 2007, the Defense Advanced Research Projects Agency (DARPA) asked American scientists to submit proposals to develop technology to create insect-cyborgs, the results of which led to a plethora of troubling and worrisome commentary.

Rather than build off of a frightening narrative that discusses the potential sinister militaristic use of such technology, this project does the opposite and imagines instead an aesthetically pleasing utopia where these insect-cyborgs have social utility and work towards the public good of humanity. Insectile indices also plays with the idea of aesthetics in our future techno-driven world by addressing whether we are more apt to silently “turn the other cheek” to more pervasive surveillance if these insect-cyborgs, or the urban landscapes they have the potential to reshape, become more aesthetically pleasing to the eye.

Keywords
+ Biomimicry
+ Internet of Things
+ Predictive Policing
+ Insect Cyborg
+ DARPA

Execution
+ Unity
+ Interactive Prototype
+ Video


The project poses the following questions: What would it be like if AI redefines natural ecosystem in a city? What is problematic with elongating the lifespan of insects through such augmentation? And, how will the life of an augmented insect differ, and interact with, naturally occurring insects?





Background
In 2007, Defense Advanced Research Projects Agency (DARPA) asked America's scientists to submit "proposals to develop technology to create insect-cyborgs". US military has been hoping to develop "micro air vehicles" – ultra-small flying robots capable of performing surveillance in dangerous territory. Cornell University researchers have succeeded in implanting electronic circuit probes into tobacco hornworms as early pupae. The hornworms pass through the chrysalis stage to mature into moths whose muscles can be controlled with the implanted electronics. Various insect species such as dragonfly, beetle, cockroaches and crickets are used for the research creating the cybernetic organism and succeeded for purpose. The lifespan of the cybug elongated through converting heat and mechanical energy that insect naturally generates.

Speculative Scenario
I imagined the future city of LA when the Cybug technology trickled down to everyday surveillance in 2027. In the future city, Cybug Farms are aggregated in the riverside which produce 70 million of cyborg insect. AI system manages quality control of the breed - classify caterpillars suitable to the system and support successful metamorphosis for ‘healthy’ cybug. The full grown moth Cybugs are freshly stored at specially designed vehicle in low temperature and safely transported to the Cybug Hotel in Los Angeles. ‘Cybug Gardeners’ supervise the process of controlling Cybug and system.

Demonstration
The demonstrations are executed with origami structure and projection mapping in order to physicalize cybug hotel. And, it describes cybug moth's movements(output) based on the input of city data collected by themselves. Virtual graphics are choreographed to describe 'Data foraging', 'Data sharing & training' and 'Synchronized behavior for swarming,' all of which investigates possible interactive movement to metaphorically visualize the cyborg AI system.

Origami structure
I choose origami structure as implication of the Cybug hotel because of the medium's interactive, transformable nature.

In the project, origami structure is a tool to help viewers to understand the cross species aspect of the cyborg moth. Unlike the moth in real nature that doesn't create habitat (they usually live in marshes and sand dune), Cybug moths reside in habitat structure with hexagonal form reminiscence of beehive.

And, moth species are selected to become a Cybug in this scenario because of their sensitivity to odors and close relation to flower. The genetic characteristics for the superpower can be easily modified for human purpose. Machine Learning models to control the cybugs are based on the Artificial Bee Colony algorithm, that makes Cybug moths to conduct efficient swarm behaviors.



Research
During the class, Duncan Wilson (Director of Intel Collaborative Research Institute for Sustainable Connected Cities) was invited to the class and talk about his recent research at London living lab: olympic park. It was particularly interesting to me when he talk about ultrasonic microphone influencing biodiversity of the area, especially bat species. Researchers created device to listen and capture the bat courses and trained data by machine learning. And he spends several years to observe bat call and tag them to embed the data set to identify what kind of bat species are found in the olympic park. His project questions how the relationship between bat and communication network interrelated, bring insight about the future of IoT and natural ecosystem, which inspires me to start the project.

Brief & Context
This project investigates new potentials for the design of Artificial Intelligence (AI) and Machine Learning (ML) for the neighborhood, human scale environment. As people traverse their day in the local urban landscape, both interior and exterior, how would pervasive AI systems manifest? We are interested in the ecologies and complex outcomes created by a range of AI embedded in the environment--what would the interactions be like, how do the different systems interact with each other, what role should design play?

We’re interested in the human and device impact of pervasive AI, and how its design will affect that impact. Will it be generic, centrally trained on massive datasets and controlled by anonymous corporations, or will the systems be locally controlled and idiosyncratic as they implement the goals of the local municipality and learn the character of the community? And what about the well being of the AI systems? Do we consider them to be productive members of the community? Should we design the ecologies to support the growth and positive influence of our AI citizens? How would that work? What are the implications of the technology and architecture of AI/ML - neural nets, supervised/unsupervised learning, the 5G edge/fog/cloud computational/network infrastructure, the methods and biases of data-scientists?

Our goal is to produce a collection of refined projects that investigate the issues around AI and ML. These can be speculative, strange and experimental, but they must engage with the real issues in the field, and must include some working elements that use or simulate AI and ML techniques. Students must be explicit about the particular areas they are exploring, and be able to discuss the insights and relevance of their work to design

Remaining questions:
+ What is the agency of the Insectile AI system? and what is the different relationship between the AI with different agents?  (such as Darpa, CIA, FBI etc.)
+ What is the possible metaphor for AI system of white collar crime? 
+ What would be the extreme utopic and dystopic scenario?
+ Who is the predator, and the audience of the cyborg insect?
+ What form of AI vandalism can be created for this scenario? 

Reference
http://www.smh.com.au/national/cybugs-a-bold-idea-hatches-20110216-1awna.html

http://spectrum.ieee.org/semiconductors/devices/nuclearpowered-transponder-for-cyborg-insect

Tools & Materials
AfterEffects, Illustrator, openFrameworks, PremierePro, Xcode

Class
Lab Projects 2: The Internet of Enlightened Things

Instructors
Phil Van Allen & Ben Hooker



Mark