Natural Language Processing: Evaluating and Improving Human Text for Future Consumption; Using Virtue Ethics to Analyze the Ethics of Amazon’s Actions Regarding Elasticsearch

My technical work and STS research have underlying connections through software, but are also connected through ethical concerns that arise in the computer science industry. More specifically, my technical project explores the usage of NLP in a large software company project to improve the readability of text and utilizes language models, a topic that has had a fair share of ethical concerns since the rise of LLMs. Thus, to gain more insight into using AI and creating software morally, my STS research examines the virtues behind the controversy between Amazon and Elasticsearch.
My technical report explores the use of NLP to evaluate text. There existed a catalog for requirements populated by forms that were not checked for readability. This led to descriptions that were confusing and hard for future users to understand. To help mitigate this, my intern team and I utilized NLP techniques to provide real-time suggestions to the user as they filled out the form to improve their writing. The suggestions consisted of grammar and spelling checks, ambiguity detection, and acronym finding. This was done using various technologies, such as Golang, Python, and a model named BERT, and was integrated into the existing catalog software’s frontend and backend.
My STS research focused on Amazon and its usage of Elasticsearch in AWS, where they released a paid version of an open-source software. To analyze this, I used virtue ethics to see if Amazon acted ethically according to specific engineering virtues and to better understand why or how certain actions they took were immoral or not. Specifically, my research paper argued that Amazon acted immorally due to failing to uphold engineering virtues of clear and informative communication, cooperation, and willingness to make compromises. This allowed for a deeper dive into making ethical decisions regarding software that could be carried over to many other situations.
There was little true overlap in the development of these two projects due to the technical project being done during an internship the summer prior. However, my STS research allowed me to gain insights that I can apply in future software that I create in my professional career. Specifically, it allowed me to have a deeper understanding of the virtues that moral engineers have and how to apply them to various situations. Thus, as I create new software in the future, I will ensure that I am aware of the people it may affect and the possible consequences that can result from its creation. Furthermore, I will view actions I take while creating through the lenses of virtues that I want to uphold as an engineer and a person, and ensure they are maintained before making any decisions.

School of Engineering and Applied Science

Bachelor of Science in Computer Science

Technical Advisor: Rosanne Vrugtman

STS Advisor: Benjamin Laugelli

Technical Team Members: N/A