AI proctoring enhances exam integrity beyond CCTV for paper-and-pen exams

AI-powered proctoring enhances exam security with real-time monitoring and anomaly detection, offering a scalable, unbiased solution beyond CCTV for both remote and in-person exams.

AI-powered proctoring is redefining exam security, offering capabilities that go beyond traditional CCTV surveillance, especially in paper-and-pen exams. Unlike CCTV, which primarily records footage for later review, AI-driven proctoring solutions use real-time monitoring, facial recognition, and advanced anomaly detection to identify potential cheating instantly. These systems not only enhance security but also ensure fairness by reducing human biases and automating key aspects of supervision.

As educational institutions and exam boards seek scalable and effective ways to uphold exam integrity, AI proctoring emerges as a modern solution, allowing for both remote and in-person exams to be conducted with heightened accuracy and impartiality.

Adarsh Sudhindra, VP of Growth & Strategy at Excelsoft, highlights how proctors enable educational institutions, exam boards, and organisations to conduct remote exams on a large scale, combining real-time monitoring, identity verification, and advanced data analysis to uphold exam standards.

Sudhindra explains that the proctor includes facial recognition, multi-camera monitoring, and real-time anomaly detection to identify suspicious behaviours, ensuring institutions can scale exam delivery securely. Additionally, the platform supports real-time communication between proctors and test-takers via text and audio chat, creating a seamless exam environment.

When asked about adaptability, Sudhindra clarifies that proctoring solutions are primarily tailored for online exams. These solutions, which include automated flagging and integration with remote assessment setups, can also support offline exams in limited-capacity test centres equipped with local server setups.

RECENT MOVE BY CBSE

Discussing the recent move by CBSE to install CCTV cameras at exam centres, Sudhindra points out that while CCTV provides real-time monitoring, AI-powered proctoring offers additional benefits.

For instance, the National Testing Agency (NTA) recently implemented AI-driven proctoring for JEE exams to prevent cheating with real-time anomaly detection. AI-powered systems offer advantages like cost-effectiveness, enhanced privacy, and scalability that traditional CCTV monitoring cannot match.

HOW AI TRANSFORMS PROCTORING

According to Sudhindra, AI reduces bias by using standardised criteria to detect behaviours, such as eye movements, that could indicate cheating, without human judgement. Additionally, flagged incidents are reviewed by humans to maintain accuracy and fairness, with a commitment to minimising demographic bias by training AI on diverse datasets.

AI proctoring goes beyond basic surveillance with specialised features like facial recognition, keystroke analysis, and behavioural analysis. Facial recognition aids in identifying faces in real-time and spotting malpractice, while keystroke analysis examines typing speed, rhythm, and pauses to identify possible unauthorised assistance.

Sudhindra notes that these tools work in tandem to ensure the right individual takes the exam and to monitor for suspicious behaviour.

AI-powered proctoring systems evaluate various data patterns to flag potential violations, achieving about 90 percent accuracy in distinguishing between legitimate and suspicious activities.

They use facial recognition, eye-tracking, environmental sounds, and browser activity to detect unusual behaviours, while body language and keystroke analysis help uncover subtle signs of cheating. To ensure fairness, flagged incidents undergo human review for an additional layer of oversight.

Data privacy is a core focus, with strict data protection protocols followed to ensure ethical AI use, particularly regarding privacy and security. Continuous research in educational AI, SaaS models, and API integrations further enhances the effectiveness of these tools.

AI proctoring also enables quicker, more accurate feedback for both students and educators by analysing exam behaviour in real time. This allows for instant feedback on conduct and detailed reports for educators, enhancing transparency and academic integrity.

Looking ahead, AI is set to play an even larger role in proctoring, with advancements in real-time object and facial detection, keystroke analysis, and anomaly detection. Active investments in research keep these solutions at the forefront of digital education security, aligning with the evolving landscape to ensure integrity and fairness in exams.