Date of Award
2025
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Computer Science
Committee Chair
Vineetha Menon
Committee Member
Letha Etzkorn
Committee Member
Tathagata Mukherjee
Committee Member
Bryan Mesmer
Committee Member
Kristin Weger
Research Advisor
Vineetha Menon
Subject(s)
Artificial intelligence--Moral and ethical aspects, Big data, Natural language processing (Computer science)
Abstract
Large Language Models (LLMs) have been a hot area for popular and research audiences. Users marvel subjectively at Artificial Intelligence (AI) outputs — so much that OpenAI delayed public release of Generative Pre-trained Transformer 2 (GPT-2) back in 2019, citing ``safety and security concerns.'' This research advances ethical and transparent AI through qualitative and quantitative measures between user intent and LLM responses in a way that aims to assist users with clear indications of good and poor alignment and of where to backtrack to minimize bias and hallucinations. The use case is recommending grant topics in a United States Department of Defense Small Business Innovation Research solicitation that match academic or industrial research teams' natural-language capability descriptions.
Recommended Citation
Bacon, Greg, "Toward measurable explainable ethical AI: an LLM-driven data analytics study" (2025). Dissertations. 455.
https://louis.uah.edu/uah-dissertations/455