Abstract
The inspiration for my group's technical project came from our own struggles dealing with lost and found systems at UVA. Losing my water bottle forced conflicts in my schedule as I spent days searching boxes stuffed with bottles scattered across numerous buildings on grounds. It felt like every class, building, and professor had their own system and nobody had definitive advice of where to look. Thus, in an effort to give back to the university community, we sought to develop a better alternative. However, modern engineering requires framing technologies as sociotechnical systems rather than purely technical artifacts, highlighting the social and professional impacts of our design choices. Because the development of this application utilized AI tools, I grew concerned regarding their long-term effect on developers like myself. This led me to investigate the impacts of AI and specific use-patterns on development efficiency, skill building, creativity, and knowledge retention for my STS research paper.
The technical portion of my thesis produced HooFoundIt, a lost and found iOS mobile app built with Swift, Xcode, JavaScript, HTML, Google APIs, and Firebase. The application centralizes recovery efforts through a 24/7 item database with verification, map and list views, posting and claiming mechanisms, and AI-assisted item classification. User testing before, during, and after our development process served as valuable evidence supporting both the ineffectiveness of the current system and the improvements brought forth with our approach. Though we were unable to deploy the system across the university community, the project serves as an important proof of concept to show that better alternatives are possible.
My STS research centered around the question: "how does the prolonged use and dependence on AI affect cognitive function and what strategies can be used to grow and retain marketable skills?" My research identified empirical evidence supporting significant cognitive offloading, decreased brain activity, creative homogeneity, and worse knowledge retention among heavy AI tool users. To understand the mechanisms behind these effects, I framed AI in software development using Thomas Hughes’s framework of Technological Momentum. This framework characterizes technological systems as building mass and progressively exerting a soft determinism on humans and other systems. By contextualizing AI as a growing system with mass stemming from huge investment, corporate and consumer adoption, and presumed efficiency gains, I presented AI reliance as a systemic push for developers to deskill rather than cases of individual laziness. Crucially, my research identified distinct use patterns which contributed to varying levels of cognitive engagement or offloading. Specific patterns like high-level planning and repeated questioning demonstrably introduced friction into the momentous system, helping to preserve brain activity, knowledge retention, and skill building.
Both of these projects are characteristic of my future in software development and the sweeping changes that AI is bringing to everyday workflows. In developing HooFoundIt and offloading some of the styling work to AI, I saw quick gains in productivity but also failed to develop stronger skills in this area, becoming reliant on AI for progressive changes. While our app was a success and showed great improvements from the current system, the partial loss of ownership with AI concerned me. My research into the cognitive impacts of AI use and the mechanisms that push developers to use it has validated many of my own feelings on the matter. Using AI sometimes feels like the easy way out, and in some ways it is. As I transition into the software engineering workforce, I feel the tension between opposing forces firsthand. While university professors warn against AI to preserve skills, the corporate world embraces it and the quick results it delivers. Coupled with the pressure of watching peers use AI to boost their own marketability, the push to take the shortcut is intense. I realize now that ethical responsibility in my own career means recognizing this momentum and choosing when to introduce friction to stay competitive without eroding my expertise.