Development of an identity validation system for online assessments
Challenge
We responded to the need of a major online university to develop a tool for validating identities in online assessments. The current pandemic situation required transitioning from in-person exams to a 100% remote setting, with online assessment processes that needed to ensure that the person taking the assessment was the correct one.
Solution
To develop a custom identity validation system based on the AWS cloud and Amazon Rekognition, allowing for entirely online assessment. Key points of the solution include identity validation through webcam, cloud architecture, integration into existing assessment processes, and reports for evaluators.
Identity validation tools for assessments rely on comparing an initial set of images taken at the beginning of the term, validated with ID cards, with various captures of the examinee's facial angles taken during the online test.
- Technology
Artificial Intelligence.
- Partner
AWS.
Result
- Reduced time spent on exam review, with only 26% of electronic proctoring reports manually reviewed (target <50%).
- Improved detection of copy attempts with a traffic light system that alerts graders to exams requiring review.
- 70,000 exams conducted using the online assessment identity validation system, involving the comparison of 9,000,000 images.