Delivery and Invigilation
Test and certify candidates 24/7, wherever they are. Rosalyn's invigilation solution keeps your test content safe and protects your candidates' privacy. Lower costs, increase satisfaction, and scale your certification program globally.
.svg)
Superhuman proctoring
A human-in-the-loop autonomous artificial intelligence invigilates the exam remotely. A single person working with AI can handle the workload of 20 human test center proctors. This human/AI collaboration makes virtual testing secure, consistent and effective.
.svg)
Human/AI Collaboration
During a testing session, the AI analyzes multiple multiplexed datastreams from the test environment for anomalous patterns. When the AI encounters a significant problem, it escalates the alert to a human for arbitration. This collaboration allows you to deliver thousands of test sessions simultaneously. You maximize security and convenience and minimize staffing and costs.
Beyond Proctoring
Invigilation is the constellation of security activities, processes, and systems of your assessment program that ensures test takers work independently and keep your test content secure. It's how you act on security threats: prevent, deter, detect, arbitrate, act, and discover. No test is perfectly secure. Continual vigilance increases the security of your exam and the value of your credential.
Using biometrics and face recognition, Rosalyn knows whether the person taking the test is who they claim to be. Impersonation and unauthorized re-taking are dramatically reduced.
A small downloadable app keeps test-takers from getting improper help from other browser windows or applications while the test is in session.
Computer vision and hearing catch gross violations in the test environment that suggest cheating. The AI invigilator continuously monitors the test environment including the computer’s operating system, webcam and microphone.
The AI watches for unauthorized people and objects like cell phones and cameras. Gaze detection can tell if candidates may be looking at proscribed aides outside the field of view.
Rosalyn’s real-time monitoring gives psychometricians invaluable data to support the validity of the assessments.
The AI analyzes the microphone input for speech suggesting improper assistance.
Answer timing and item scores are datastreams the AI uses to separate innocuous behavior from potential violations.
Profile, location and test history add powerful dimensions to the AIs invigilation model.
An advanced rules engine operates on the AI output to trigger interventions and escalations. Interventions can range from direct warnings to instant referral to a human arbiter..
The AI is constantly learning from an enormous dataset of past sessions to better spot subtle cheating cues. (invisible to any human)
Learn more about Rosalyn
.svg)