V. Methods of Proctoring
d. Automated Online Proctoring with AI
Automated online proctoring with AI uses artificial intelligence to monitor students during online exams, replacing human proctors. AI algorithms assess student behavior, including eye movements and screen activity, to detect any suspicious activities. The advantages of AI-powered proctoring include scalability, reduced bias, immediate feedback, and cost-efficiency, while best practices for implementing this technology include providing software familiarization and clear guidelines for students.
BASIC Information
Automated online proctoring with AI involves using artificial intelligence to monitor students as they take exams online. Instead of human proctors, sophisticated algorithms assess the student's behavior during the exam to detect any suspicious activities. For instance, the AI can monitor the student's eye movements, screen activity, and even scan the room for any unauthorized materials or devices.
READ
A Systematic Review on AI-based Proctoring Systems: Past, Present and FutureIf a student is taking a multiple-choice exam on diagnosing illnesses, the AI will ensure they don't look away from their screen frequently or access unauthorized resources.
When students answer questions about patient care, the AI can monitor for signs like frequent tab-switching that might indicate they're searching for answers online.
INTERMEDIATE Information
Advantages:
Scalability: AI can monitor numerous students simultaneously without human intervention, making it easier to handle large numbers.
Reduced Bias: AI doesn't have human biases, ensuring consistent monitoring for all students.
Immediate Feedback: AI can immediately flag suspicious behaviors, giving institutions the chance to act promptly.
Cost-Efficient: Over time, AI can be more cost-effective than continuously hiring human proctors.
False Positives: AI might sometimes flag benign behaviors as suspicious.
Technical Issues: Malfunctions or software errors can disrupt the examination process.
Privacy Concerns: Students might be uncomfortable with being monitored by AI, fearing misuse of data.
AI proctoring can efficiently handle theoretical exams where students answer questions about medical procedures, diagnosis, or drug interactions.
However, for practical or oral exams, where students might need to demonstrate a skill or explain a concept verbally, AI's capacity to evaluate might be limited.
ADVANCED Information
Pre-exam Preparations:
Software Familiarization: Provide students with opportunities to familiarize themselves with the proctoring software before the actual exam.
Clear Guidelines: Offer comprehensive instructions about what constitutes suspicious behavior. For instance, looking away to think might be flagged, so students should be made aware.
Instant Notifications: If possible, set the AI to notify students instantly if a potentially suspicious activity is detected. This way, they can correct their behavior immediately.
Post-Exam Review:
Human Oversight: Even with AI, it's essential to have human reviewers check flagged incidents to discern false positives.
Feedback Loop: If students raise concerns about their flagged behaviors, take them into consideration to refine the AI for future exams.
Transparency: Clearly communicate how data is collected, used, stored, and when it will be deleted.
Data Security: Ensure robust encryption and secure storage of recorded data.
Bias Checking: Regularly test and update the AI to ensure it doesn't inadvertently develop biases against certain groups of students.