Abstract
My technical work and my STS research are connected through their shared focus on artificial intelligence and how it shapes user interaction. While both projects examine AI systems, they approach the topic from different perspectives. My technical project focuses on improving how users access and understand information through AI, while my STS research explores the ethical implications of AI systems that simulate human interaction. Together, these projects examine not only what AI systems can do, but also how their design influences user behavior and responsibility.
My technical project, PRAGUVA (Personalizable RAG for UVA), is a graph-based Retrieval-Augmented Generation system designed to help students navigate the University of Virginia’s academic network. The system allows users to ask questions and receive synthesized responses generated from a knowledge graph of professors, courses, and research topics. Unlike traditional search tools, PRAGUVA emphasizes transparency by visualizing the connections used to generate each response, allowing users to explore how the system reached its answer. It also incorporates personalization by using user-specific academic data, such as transcripts, to enhance responses. The goal of this project is to create a more intuitive and trustworthy academic search tool that allows students to better understand and utilize available resources.
My STS research examines character-based AI chatbots and the ethical concerns surrounding their design, particularly in emotionally immersive interactions. Using the User Configuration framework, my paper argues that these systems are designed with assumptions about users’ emotional autonomy, the idea that users can safely engage with simulated intimacy without harm. However, this assumption does not hold true for all users, especially adolescents, who may interpret chatbot interactions as genuine emotional relationships. As a result, psychological harm becomes a foreseeable outcome of design rather than simply a result of misuse. My research also explores how companies respond to such harm through policy changes, often focusing on safeguards rather than reconsidering the underlying assumptions embedded in the technology.
Working on these two projects together added significant value to both. Developing PRAGUVA made me think carefully about how users interact with AI systems and how design choices can improve trust and understanding, while my STS research encouraged me to critically examine the assumptions embedded in AI design, particularly how systems may overestimate user capability or awareness. This perspective emphasized the importance of transparency in PRAGUVA, as making the system’s reasoning visible helps prevent users from blindly trusting its outputs. Overall, working on both projects allowed me to approach AI from both a technical and ethical perspective, as my technical work focused on building a system that is useful and interpretable, while my STS research highlighted the broader consequences of design decisions. Together, they demonstrate that engineering is not only about creating effective technologies, but also about understanding how those technologies shape human behavior and responsibility.