Abstract
The technical project that I worked on consisted of creating a full stack website in my Advanced Software Development class. My assigned group and I developed code for the front end, back end, and database for a whistleblower site, called Date Warner, designed for users to go in and review and see reviews of others on dating sites. This was designed to protect people on the sites from potentially harmful users. My STS research project, although not directly related to the site, focuses on ensuring that technology is used for the good of the population, and people that use this technology are not just taken advantage of. My STS research project dives into the use of Machine Learning in hiring practices, uncovers the bias present in these algorithms, and suggests ways to mitigate this inherent bias. My group was motivated to build Date Warner after we were tasked with building a whistle blower application. We wanted to create a site that would be relevant to a great population of college kids, and something with the purpose of keeping others safe- using technology for good. As for the motivation behind my STS research, we are at a point in our college careers where it is imperative that we apply to jobs to launch our future careers. The hundreds of applications that students are turning in are met with a significant amount of rejections, often with little to no reason on why we were turned away so quickly. The answer to this in many cases is that our resumes are being sent through a black box algorithm that analyzes them and evaluates whether each applicant is good for the job. However, many of these Machine Learning algorithms are developed without ethical consideration, and inherently biased data slips into the algorithms, which is ultimately being reflected in their outputs.
The technical portion of my research consisted of working with a team to develop a fully functional website, leveraging the tools of React, Django, AWS storage, and Heroku. My position on this team was as scrum master, where I directed my team in the development. I ensured we were on track through all stages of building the site, and allocated portions of the code for each team member to work on. My primary role in the development sector came in two forms- connecting the back end logic to the front end, and developing an engaging front end design for users. In connecting the back end to the front end, I wrote code that ensured all of the data that was being entered by the users was smoothly transferred to the backend for storing and future recovery purposes. As each user created a new entry of a review of another user on a dating app, the code that I wrote ensured that the proper information was grabbed from the entry stored in the proper place in the database. This process consisted of extensive research and testing to ensure its smooth execution. As for developing the front end, I went through the process of trial and development, testing with potential users and stakeholders on which designs would be the most engaging for a user. This process of development helped to teach me the steps that go into creating a full working website. Although this site is only available for class use, the idea and potential behind it could help to ensure safety for dating site users.
As for my STS research, I did a deep dive on the bias present in Machine Learning hiring algorithms, and proposed several solutions to mitigate this bias as much as possible. I began by researching potential causes of the issue, which include flawed training data and lack of ethical intervention at every step of building the system. I proceeded to generate a potential fix to the problem. I used Value Sensitive design as my base framework for my contribution, which was to ensure that the ideas of ethicists and social scientists were considered in the development process of machine learning algorithms that are being used in the hiring process. In an attempt to create a truly fair algorithm, it would be ignorant to rely solely on the brains of engineers, as their specialties lie in the technical development field. They often cannot evaluate the fairness metrics of an algorithm throughout the development of it.
Throughout my technical project and STS research, I gained insights on the importance of considering ethical standards in engineering work. By doing this, engineers are able to build systems that can be trusted by the public. A system that is built with the intention of overall success in multiple aspects of life is one that is built with diligence, and people can rely on them without always having to consider the risks that they might ensue.
