Georgia Institute of Technology
Recommendations for an Artificial Intelligence Policy for International Nuclear Safeguards: Three Public Policy Approaches
Abstract
dc:description.abstractIn the present research, I incorporate three public policy approaches to define recommendations for an AI policy for international nuclear safeguards. First, I provide an overview of international nuclear safeguards and potential interest in emerging technologies. In Chapter 2, I use an approach inspired by the policy diffusion literature, which describes how actors learn from and adopt policies from those with whom they have frequent interactions. Using the policy diffusion lens, I assess the AI policies of actors within the policy environment of the IAEA’s environment. I use two methods of natural language processing to understand clusters, themes, lexical alignment, and temporal trends of AI policies that could diffuse to the IAEA. In Chapter 3, I use a public values approach to understand priorities of AI public values and the performance of international safeguards scenarios on those values. I describe an experimental survey used to assess opinions of international nuclear safeguards experts. In Chapter 4, I use a responsible research and innovation approach that recommends codes of conduct to fill the gap until formal policies are enacted. I use a systematic literature review to collect 682 labeled recommendations in a dataset that I call the Responsible AI Recommendations (or RAIR) dataset. Finally, in Chapter 5, I compile recommendations from the three public policy approaches into a Model AI Policy for international nuclear safeguards.
Degree
thesis:*- Name thesis:degree_name
- Public Policy, PhD
- Department thesis:degree_department
- Public Policy
- Grantor
- Georgia Institute of Technology
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Gastelum, Zoe N.
- Advisor dc:contributor.advisor
-
- Sugimoto, Cassidy R.
- Committee members dc:contributor.committeemember
-
- Kingsley, Gordon
- Biddle, Justin
- Kosal, Margaret
- David Stracuzzi
- ORCID person:identifier.orcid
Rights
- Language dc:language.iso
- English
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/81709