Privacy in Georgia Tech Research
SATURN is one of the more recent projects developed and published by the Georgia Tech Ubiquitous Computing lab here on campus. It’s a thin, flexible, paper-like material that can be manufactured in large scale from cheap parts, and functionally act as a microphone surface that can be deployed on any surface to catch ambient sound waves. The materials are arranged in such a way that this collection of vibrations can be done with no battery; the waves actually manipulate the material in such a way that power is created in the sensor. The most recent development in this project is implementing a small antenna and transistor circuit so that this microphone material can wirelessly transport its collected audio data to an external processing unit. Imagine someone walking up to your desk, slapping a post-it note on the corner and walking away without you realizing. If the project succeeds in its vision, then this post-it note can hear everything you say without anyone being aware of it, and this character can record and store all this audio through their own machine.
I work in this lab on this very project conducting experiments with the antenna performance, and there are a team of students working on similar projects for ubiquitous sensing at Georgia Tech as well. If this class has made me think about anything more skeptically, it’s the balance between usability and security, and learning about sacrifices of usability for the sake of privacy has made me reflect much more on the implications of this work. Originally, coming purely from a design and naive application perspective, I thought that SATURN would have an incredible amount of purely good use if deployed in the environment. Imagine being able to truly have a smart home that can be controlled with voice commands no matter where you are in the house, or imagine the case where this microphone material is deployed near a street and can pick up and pinpoint locations of car wrecks on busy roadways. Smart cities could be developed to measure traffic flow and civilian activity through recording audio throughout a city, and there are so much personal and social good that can come out of these applications. This leads me to think that SATURN is an incredibly usable technology that has an absolutely huge amount of theoretical scenarios, and before taking this class I didn’t think more critically about the implications of such usability.
As Dr. Das has said many times, usability and privacy is a sliding scale with almost unavoidable tradeoffs between the two and user priorities that are constantly moving around. Since SATURN is such a (theoretically) usable technology in that it’s ambient and ubiquitous, it’s also trading a ton of privacy in that it’s ambient and ubiquitous. For every good use case I can think of, I could imagine a nefarious use as well. The more I reflect on these issues, the more I realize that it this two-sidedness could be a unavoidable result of usability. Is it possible to have the best of both worlds? There’s not an easy answer here, but it’s a very interesting question.
For optimistic purposes, I tried to perform some research into privacy in ubiquitous computing and came across an interesting article that may have been described in the class; ‘Cerberus: A Context Aware Security Scheme for Smart Spaces’. The publication proposes some system and conceptual architecture to bring security into the field of ubiquitous computing, and most of the ideas here don’t really apply but one that stuck out to me: the emphasis on transparency in ubiquitous systems. This is harder in software system design because you have to take care to let the user know what’s going on in the background in an understandable and non-overwhelming way. In this scenario with physical sensors, however, we can be transparent by simply designing this material to be clearly differentiated from its surroundings. For SATURN, transparency is thrown away if we design this material to look exactly like any other piece of paper because nobody is aware of the technology’s presence or activation. Transparency is respected if we even do something as simple as painting a big intimidating skull on every square foot of material. Sauvik has said that in the online world, it’s hard to present digital threats with the same instinctual danger as a robber coming at you with a knife or a bomb in front of you. When dealing with hardware devices that will physically be in the same room as the user, I think we can actually strive to bring forth that natural sense of danger or apprehension. This physical design challenge to avoid a lack of transparency is something I’d like to explore in this project further, and requires a lot more design ideation than printing a big skull on everything.