You are now in the main content area

Virtual proctoring tools available at Toronto Metropolitan University

What is automated virtual proctoring?

A automated virtual proctoring system uses artificial intelligent technology to monitor students while they are taking their exams in the location of their choice, usually their homes. The system uses web cameras and screen monitoring software. It includes an identity verification process using the student’s image and their photo ID that are captured where the instructor can review them. The system also records exam sessions and provides flags in the recording where the student behaviour matches the list of academic misconduct either provided by the system or the instructor.

At Toronto Metropolitan University, we have Respondus LockDown Browser + Respondus Monitor available for automated virtual proctoring.

 

Respondus Monitor

Respondus Monitor (external link)  is a fully automated virtual proctoring system. It is accessed seamlessly through the D2L Brightspace Quiz tool, therefore students and instructors do not need to create separate accounts in Respondus. Respondus Monitor enables students to take online exams at any time according to the D2L Brightspace Quiz availability.

The Respondus application, which is a browser (Respondus LockDown Browser (external link) ), will lock the student in the browser and the student will be unable to access anything else on their computer which the instructor has listed as prohibited. There is also an allowed list of applications and URLs the instructor can configure.

A webcam is used to record themselves during the exam. The video recordings are streamed to the Respondus server during the exam and stored on the Respondus servers afterwards. A "startup sequence" guides students through requirements selected by the instructor, such as showing identification or making a short video of the exam environment. Only the instructor and TA/GA of the course can review video recordings and the results of the proctoring session. The videos are not downloadable and cannot be easily shared with others. 

Facial detection and artificial intelligent algorithms are being used to analyze the recordings and flag certain behaviours during an exam session. Video segments with potential exam violations are flagged for the instructor review.

Virtual proctoring feature

Respondus LockDown Browser + Respondus Monitor

Integration with D2L Brightspace

Yes

Exam preparation

Using Brightspace Quizzes

Students need to install the application.

Yes

Student’s ID verification process

Instructor has the option to require (or not) the following:

  • Student photo
  • ID card
  • Environment check

Students taking the exam



Exam which is in the Brightspace Quiz can be taken using a desktop computer or iPad

Instructors have the option to disable the following:

  • Print
  • copy/paste
  • Screen captures 
  • and have a list of allowed links/websites and applications.

Students video recording while taking the exam

The video is streamed to and recorded in the Respondus server.

  • Internet outage will cause a failure in the video recording. 
  • Internet disruption, e.g. reduced bandwidth, will not cause failure. Respondus will adjust.

Students submitting the exam

Submit the D2L Brightspace Quiz in Brightspace, logout from Brightspace, and close the browser.

Instructors reviewing the result.




Instructors are presented with “flags” as noted by the AI.

Instructors need to review the flags and discern whether they are academic misconduct or not. 

Flags

Flagged Events (external link) :

  • Missing from Frame
  • Different person in Frame
  • Multiple persons in Frame 
  • An Internet interruption occurred
  • Video frame rate lowered due to quality of internet connection
  • Student exited LockDown Browser early
  • Low Facial Detection
  • A webcam was disconnected
  • A webcam was connected 
  • An attempt was made to switch to another screen or application
  • Video session terminated early
  • Failed Facial Detection Check
  • Student turned off facial detection alerts