AI & Inequities

The following is an excerpt from “AI & Inequities in the Hiring Process for Black Educators” by Dana Dyer L’22, produced as part of Senior Adjunct Professor of Global Leadership Rangita de Silva de Alwis’s “Policy Lab: AI and Implicit Bias” at the University of Pennsylvania Carey Law School:

In the classroom, diverse representation among staff and students is critical for an enriched education experience. Unfortunately, schools across the country significantly lack this diversity. The National Center for Education Statistics found that Black teachers make up less than 7% of the teaching force in public schools and about 10% in charter schools, despite evidence of the positive impact Black teachers have on Black students’ test scores and graduation rates.

In Philadelphia, there are only 23% of Black teachers in public and charter Franchessca Dyer, English Department Head, English Facilitator, and Induction Coordinator at Sankofa Freedom Academy Charter School in the Kensington section of Philadelphia.schools, while over 54% of the student enrollment are Black students. The teacher force also does not represent Philadelphia’s demographics as 35% of Philadelphians are white residents and 44% are Black residents. Philadelphia educator, Franchessca Dyer (pictured to the right) told the author in an interview, “I absolutely love teaching. There’s something about the feeling I get when students who I have taught to learn, excel, and move on to do great things. I like to think that I played a part in that; shaping young people for the future.” The hiring process for educators must prioritize retention of diverse teachers from various racial and ethnic backgrounds.

Hiring platforms, like Indeed and LinkedIn, help employers hire candidates who match the listed job description to their listed qualifications using artificial intelligence (“AI”). This expedites resume review down to milliseconds. However, the expedited process does not eliminate discrimination or bias Black educators encounter in the hiring process. In fact, bias may be rooted in the data that trains artificial agents to find patterns. A Harvard Business Review article states, “[a]lgorithms are, in part, our opinions embedded in code. They reflect human biases and prejudices that lead to machine learning mistakes and misrepresentations.” AI is a civil rights issue – left unmonitored, AI can perpetuate the same inequities in hiring we wish to improve.

This report explores potential algorithmic biases at the intersection of race, gender, and education. With the use of quantitative and qualitative data, the author tells the stories of Black educators who experienced algorithmic bias in the hiring process or other biases throughout their respective careers in education. The author divides the report in three sections to reflect the most alarming data: Transparency in AI, Stereotype Threat, and the Normalization of AI Bias. This single seminal report cannot address the plethora of inequities Black educators face in hiring and the workplace, nor does this report serve to address each potential issue in AI. Instead, this report aims to alert users and vendors of algorithmic bias and to consider more transparency and accountability before reliance on an AI-based platform.

In Professor de Silva de Alwis’s Policy Lab, students were joined by brilliant scholars and revolutionary leaders in AI tying together interdisciplinary concepts from the Global North and South. This report will include thoughts and remarks about AI bias from several speakers to further support the data collected from Black educators.

The following guest speakers highlighted:

• • Judith Donath – Author of The Social Machine, writer, designer, and artist
  • Ethan Zuckerman – Social scientist and Philosopher, former Director of the Center for Civil Media at MIT, former Associate Professor of the Practice at the MIT Media Lab
  • Heather Sussman – Partner and head of Orrick’s Global, Cyber, Privacy & Data Innovation Group
  • Deborah Raji – computer scientist, worked closely with Dr. Joy Buolamwini in Algorithmic Justice League, MIT’s 35 under 35 innovators

Read Dyer’s full report.

Read “Policy Lab on AI and Implicit Bias” by Veda Handa LLM’22, also produced as part of the course.