Equity Implications of Using AI Tools in the College Admissions Process

Julie M. Smith, Jordan Williamson, Monica McGill

Track: Responsible AI for Education (Day 2)

Paper Length: short-paper (2 pages + references)

Keywords: equity, college admissions, AI

TL;DR: Based on an algorithmic auditing simulation of those working in college admissions, we present a framework for ensuring that AI tools used in the admissions process are equitably used.

Abstract:

College admissions offices face the complex task of fairly
evaluating candidates for admission, including students from
historically marginalized backgrounds. This task is complicated by decreasing budgets, the end of affirmative action,
and, in some states, legal restrictions of diversity, equity, and
inclusion efforts. The use of AI tools in the admissions process presents both an opportunity and a challenge. In this poster, we describe a brief experiment in using a LLM for a simulated college admissions task, alongside recommendations for ways to use AI tools to promote equity in the admissions process.

Submission Number: 39