Gryzbowicz, Paulina

The Social Impact of Algorithmic Bias

Paulina Grzybowicz is a Honors student at DePaul University pursuing a Computer Science degree. Due to graduate in Fall 2020, she is currently interning at Oracle in Chicago. Outside of programming, she serves as a CQM in DePaul’s Chicago Quarter program and is an active member of the Polish-American community in Chicago. Her passion for social justice is what inspired her thesis topic, and she hopes to use her skill set to promote social good in the future.

Major: Computer Science

Senior, College of Computing and Digital Media

---

Abstract

Algorithmic biases are the errors, tendencies, and trends within an Artificial Intelligence (AI) model that have the potential to create unfair or discriminatory outcomes. There are many factors that influence the emergence of bias in such a system, including how the model is designed, how training data is gathered and used, and pre-existing individual and institutionalized biases. The widespread use of predictive AI models (e.g political campaigns, financial institutions, Internet search engines, etc.) means that any negative bias that emerges can have a serious impact. Algorithmic bias can discriminate along gender, race, and ethnicity categories and work to perpetuate already existing injustices in the U.S. This paper explores how algorithmic biases are formed, their social implications, and the attempts being made to mitigate them.

Thesis Director: Peter Hastings

Department: Computer Science

Faculty Reader: Martha Martinez-Firestone

Department: Sociology, Critical Ethnic Studies

---

Project Poster & Narration