Abstract
Deepfake Detection, the Safety of Next Generation Rocketry, and Utilitarianism
Relationship between technical and STS research
At a glance these projects are fully disjoint, although a more careful analysis of their topics reveals a small intersection. My technical project explored methods of creating deepfake detectors designed to be more accessible to the public, with the hope of being able to provide an easier to understand way for average people to interact with the media they see on a daily basis. This goal inevitably brings into question the ethics of trying to detect deepfakes, which is where we begin to see an intersection with my STS research. The analysis I provide as to whether or not Starship can be an ethical construction is drawn directly from a utilitarian viewpoint. As the volume of cargo and passengers it can reliably and safely ferry increases, the significance of it being unable to handle catastrophic failure decreases. This is similar to how it is generally accepted that a major airline failure has a high probability of losing most, if not all, of its passengers. Deepfakes make us wonder if it’s ethical to tell people what is real. Is the benefit to society of spotting fake media worth the risk of false positives? These questions, as with the question of the ethics used in Starship’s design, can all be answered in Utilitarian fashion.
Technical Research - Deepfake Detection
In my Technical Report, I discuss a novel method of deepfake detection which is created by combining individual metrics, or algorithms, capable of detecting fake content. The motivation for such a method comes from the current state of the deepfake detection industry. Presently, large companies are pioneering the software to do this detection, however their methods typically require extensive computational resources and make use of complex algorithms that render an understanding of the system out of reach for most people. My project explores a) whether one can combine two simple detectors and see a net gain in accuracy, and b) whether a simple interface can be created which makes the detection process accessible to people with little to no computer experience.
STS Research - Safety of next-generation rocketry
In my STS Research Paper, I explore SpaceX’s new Starship launch vehicle and the fact that it currently lacks a launch escape system. Up until Starship, every spacecraft intended for humans has been designed from the beginning with some escape technology in mind. As a result, my research is into whether or not the Starship can be considered an ethical construction. To analyze the situation, I use the Utilitarianism ethical framework and argue in favor of Starship using real facts and figures from similar SpaceX ventures to estimate safety, risk, and utility of the entire Starship launch architecture.
Joint project value
Working on these projects during the same semester provided additional opportunities for reflection during the planning and writing of my STS Research Paper. Although I finished the technical report before beginning work on the analysis of Starship, starting that analysis with the thoughts of deepfake ethics already swirling in my mind is what prompted me to employ Utilitarianism. While writing the Technical Report, I thought many times about the benefits deepfake detection can offer society, but also the potential harm a reputable detector could cause if it were to misjudge a piece of content. It turned out that, for me, weighing these rewards and risks was ethically similar to weighing the different arguments as to whether a launch escape system ought to be necessary or not. Consequently, the overarching ideas in my approach to deepfake ethics proved to be similar enough to Starship’s escape system dilemma to warrant the same general framework being used in the thinking behind both projects.
