Cybersecurity of Online Proctoring Systems

Abstract

The online proctored examinations are adopted exceedingly in all forms of academic education and professional training. AI with Machine Learning technology take the leading role in supporting authentication, authorization, and operational control of proctored online examination. The paper discusses how administrative, physical, and technical controls can help mitigate related cybersecurity vulnerabilities of online proctoring systems (OPS). The paper considers two classes of OPS: fully automated AI-enabled systems and hybrid systems (automated AI-enabled with an expert live proctor in control). Based on the review of 20 online proctoring systems, the paper discusses methods and techniques of multi-factor authentication and authorizations, including the use of challenge-response, biometrics (face and voice recognition), and blockchain technology. The discussion of operational controls includes the use of lockdown browsers, webcam detection of behavioral signs of fraud, endpoint security, VPN and VM, screen-sharing and keyboard listening programs, technical controls to mitigate the absence of spatial (physical area) controls, compliance with regulations (GDPR), etc. Other topics discussed include confidentiality of the exam content, logging of control data, video and sound recording for auditing, limitations of endpoint-based security protection and detection techniques of behavior-based cheating and the effect of new intrusive technology on students’ privacy. In conclusion, the paper lists advanced features of online proctoring systems.

Ähnliche Arbeiten