Analyzing Unconscious Bias in Indeed’s Employee Resume Search
| dc.contributor.advisor | Sin, Steve | |
| dc.contributor.author | Antony, Rachel | |
| dc.contributor.author | Bral, Benjamin | |
| dc.contributor.author | Gleason, Seth | |
| dc.contributor.author | Ihm, Joanna | |
| dc.contributor.author | Malhotra, Aarushi | |
| dc.contributor.author | Mathew, Philip | |
| dc.contributor.author | Nagaokar, Soham | |
| dc.contributor.author | Rajala, Johnny | |
| dc.contributor.author | Truong, Kyle | |
| dc.contributor.author | Zhu, Daniel | |
| dc.date.accessioned | 2024-03-12T15:38:50Z | |
| dc.date.available | 2024-03-12T15:38:50Z | |
| dc.date.issued | 2023 | |
| dc.description.abstract | This project analyzes if artificial intelligence (AI) hiring systems demonstrate racial bias as measured by prestige bias against graduates of historically black colleges and universities (HBCUs), measured in a number of different metrics, and how that bias can be mitigated. The metrics we used to measure prestige were: size, university rankings, public vs. private universities, and attendance of HBCUs vs. attendance of non-HBCUs. We examined how the hiring site Indeed utilizes AI to list candidate resumes by relevance and measured the relationship between candidates’ resume rankings and the universities they attended. While we found no significant difference between the overall average rankings of applicants from HBCUs and applicants from non-HBCUs, we did find significant differences between these applicants when we made comparisons based on variables such as major, experience level, and most recent company size. Future research on this topic includes training an AI model on the collected resumes to see if the same results are generated and adjusting the model to mitigate said biases. This research will shed light on the bias embedded in human hiring departments. With businesses considering AI as a tool for hiring, companies must understand that AI hiring systems can perpetuate the same biases found in human hiring on a larger scale. | |
| dc.identifier | https://doi.org/10.13016/tmqp-ypve | |
| dc.identifier.uri | http://hdl.handle.net/1903/32359 | |
| dc.relation.isAvailableAt | Digital Repository at the University of Maryland | |
| dc.relation.isAvailableAt | Gemstone Program, University of Maryland (College Park, Md) | |
| dc.subject | Gemstone Team DeBIAS | |
| dc.title | Analyzing Unconscious Bias in Indeed’s Employee Resume Search | |
| dc.type | Thesis | |
| local.equitableAccessSubmission | No |