Abstract
This Master of Public Administration capstone investigates the structural and institutional constraints affecting artificial intelligence development in the United States and evaluates congressional policy responses. The study examines interdependencies among workforce preparedness, computing capacity, energy infrastructure, capital investment, and regulatory fragmentation through a multi-method research design.
LLM-assisted qualitative content analysis of more than 1,000 secondary sources informs the identification and classification of development bottlenecks. A Bottleneck Composite Index integrates frequency, impact, resolution, and urgency to establish comparative priorities. Decision Making under Deep Uncertainty (DMDU) then evaluates candidate policy strategies across 10,000 simulated futures, incorporating Robust Decision Making, Dynamic Adaptive Policy Pathways, and exploratory modeling. The analysis compares policy performance, institutional feasibility, and adaptability under alternative modeled conditions to inform recommendations concerning workforce development, computational access, energy infrastructure, capital allocation, and regulatory coordination.
Analytical contribution
I led the project’s Decision Making under Deep Uncertainty (DMDU) analysis, applying software development and computational modeling to the evaluation of public-policy alternatives. This work translated policy questions and evidence-derived bottlenecks into a simulation framework for examining the comparative performance of candidate interventions under uncertain future conditions.
The analysis evaluated strategies across 10,000 simulated futures, integrating static robustness assessment, adaptive policy selection, and exploratory analysis of scenario-level outcomes. My contribution connected computational methods with public-administration concerns, including institutional feasibility, implementation constraints, and policy adaptability. The simulations compare strategies within the study’s specified assumptions and scenario ranges.
Research questions
- Which structural and institutional constraints recur across the literature on American AI development?
- How can the frequency, impact, resolution, and urgency of these constraints be operationalized to establish comparative policy priorities?
- How do alternative congressional policy strategies perform across plausible future conditions, accounting for mitigation effectiveness, institutional feasibility, and adaptive capacity?
Research design and methods
- Qualitative evidence synthesis: LLM-assisted content analysis using an iteratively refined codebook, human review, and coding comparisons across secondary sources from industry, government, think tanks, academia, and media.
- Composite-index construction: A Bottleneck Composite Index combining coded frequency, impact, resolution, and urgency to rank identified constraints and inform candidate policy interventions.
- Robust Decision Making: Comparison of fixed policy strategies across 10,000 simulated futures with varying bottleneck severity, incorporating institutional and political feasibility into the assessment.
- Dynamic Adaptive Policy Pathways: Threshold-based selection among candidate strategies as modeled bottleneck conditions change.
- Exploratory modeling: Analysis of scenario-level outcomes to identify conditions associated with stronger or weaker policy performance.
Authorship and academic context
Coauthors: Garrett Collier, Jarred Cornell, Sherrelle Filimoehala, and Luis Rubalcava.
Program: Master of Public Administration, University of Southern California, 2025.
Faculty advisor: Logan O’Shaughnessy.
This coauthored capstone report was completed in August 2025. The complete document includes the research methodology, analytical findings, policy recommendations, references, and supporting appendices.
Read the complete capstone report (PDF, 91 pages)
Academic foundation
My education brings together a Bachelor of Arts in Computer Science from Hendrix College (2022) and a Master of Public Administration from the University of Southern California (2025). The capstone integrates these disciplines through computational policy analysis, institutional assessment, and evaluation of public-sector responses to technological change.