Tanusree Sharma's Dissertation Proposal Defense

Her committee includes Associate Professor Yang Wang (chair), Associate Professor Yun Huang, Andrew Miller (ECE), and Sauvik Das (HCI Institute, Carnegie Mellon).

Abstract:

A major criticism of past AI development is the absence of thorough documentation and traceability in design and decision-making, leading to adverse outcomes such as discrimination, lack of inclusivity and representation, and breaches of legal regulations. Underserved populations, in particular, are disproportionately affected by these design decisions. Conventional law and policymaking methods have constraints in the digital age while traditional methods like interviews, surveys, and focus groups for understanding user needs, and expectations have inherent limitations, including a lack of consensus and regular insights. In my proposal, I aim to address these challenges by facilitating the decision-making process where users can actively participate in determining the level of personalization, they would like to experience in AI systems.

This involves introducing technical interventions, equipped with decentralized mechanisms, particularly Decentralized Autonomous Organizations (DAOs) mechanisms that incorporate various consensus-building approaches, enabling end-user participation in the decision-making process for AI rule development. Additionally, I also introduce a novel method for social alignment that ensures transparency and traceability in AI governance. The insights from this proposal have practical implications not only for democratic design in AI but also for broader governance contexts with centralized structures.