Report on ExamSoft’s ExamID Feature (and a Method to Bypass It)
Journal of Robotics, Artificial Intelligence, and Law (2021, Forthcoming)
6 Pages Posted: 8 Jan 2021 Last revised: 10 Mar 2021
Date Written: 2021
Abstract
As a result of the global COVID-19 pandemic, many academic institutions have been forced to move their courses online. This has necessitated schools to implement new technologies to remotely administer examinations. One of the most prominent vendors offering software to allow for this is ExamSoft. Some institutions have implemented an additional authentication feature from ExamSoft called ExamID. To confirm students’ identities, ExamID uses a form of artificial intelligence - facial recognition technology - to match an image of a student in its database to an image the student takes immediately before an exam. In theory, ExamID then matches the stored image with a second image as a means to confirm that the correct person is taking the exam. Unfortunately, just as has been the case with other efforts to use facial recognition technology, ExamID has been alleged to be error-prone for students of color attempting to use it. ExamSoft contests these claims.
Regardless of the accuracy of the public allegations against this software, it still puts academic institutions in a position where responding is important. Irrespective of whether ExamID negatively impacts an individual student of color directly, it is vital to students’ well-being and sense of inclusion that the entire community can trust that their school is assuring vendor products treat all students equitably.
Suggested Citation: Suggested Citation