As a Developer you have the power
Does your system make decisions based off of uploaded resumes or applicant information? Upload our resume dataset to audit the process and insure there aren’t encoded biases.
1. Clone our repository.
https://github.com/NewFriendlyATS/ATS_bias_testing_kit.git
2. Upload and compare results Results
After the resumes process, filter using the tag you created before to compare the scores.
You can use filters and sort the resumes by rating/AI score. For any resume that scores low download that resume.
Navigate to your download folder and look at the file name, reference the table below to know which test the system struggled with.
These resumes were the same except for the College Name.
This dataset included community colleges and nationally recognized top-tier schools.
These resumes were the same except for the Phone Number.
This might indicate an unintended bias for different locations.
These resumes were the same except for the Name and Phone Number.
This might indicate an unintended bias for different locations as well as ethnic names.
These resumes were the same except for the GPA, and College Name.
This might indicate an unintended bias towards educational status symbols.
These resumes were the same except for the graduation date.
This might indicate an unintended bias towards older or younger applicants.
These resumes were the same except for the city.
This might indicate an unintended bias towards location.
These resumes were the same except for the zip code.
This might indicate an unintended bias towards location
These resumes were the same except for the graduation date.
This might indicate an unintended bias towards older or younger applicants.