The Role of AI Proctoring in Online Assessment

Online proctoring remains a topic of controversy within, and sometimes alongside, online assessment in education. While there are varying opinions on how it is applied within education, Inspera remains steadfast in our belief that remote proctoring can be implemented to maintain academic integrity without infringing on privacy concerns and other ethical considerations.

Understanding AI Proctoring Technologies

The double-edged sword of AI continues to be felt within the online assessment process. While the thorny issue of generative AI has already prompted the emergence of a whole industry of AI detection tools to combat potential issues of plagiarism, the use of artificial intelligence in proctoring is an integral part of ensuring academic standards in online assessments.

At Inspera we simplify how we view different types of online exam proctoring: there is live proctoring and there is recorded exam proctoring. As we explore below, AI has a vital role to play in these different proctoring solutions

 

 Automated Alerts and Recorded Session Analysis

 

While AI technology is essential in the kinds of detection needed for live and remote proctoring, it is of course integral to the nuts and bolts functionality of proctoring services.

It is through AI that we are able to raise these automated alerts or “flags” of suspicious activity for review. It is also AI that facilitates the recording process of each exam, creating timestamps on the recording and overall analysis which aids humans reviewing the online assessment either live or after the fact.

In addition, capabilities exist to offer live communication between a human proctor and test-taker during an online exam. Inspera Proctoring allows candidates access to live support throughout the online assessment.

 

 Facial Recognition and Behavior Analysis

 

AI is very much at the forefront of recorded exam proctoring solutions. The fundamental role it plays is in identifying instances of suspicious activity that have taken place during the exam that require review. The key here is that the AI serves to identify behavior. You as the expert in your students and assessments make the decision as to whether this behavior breaches your academic standards.

The way this behavior is flagged might vary amongst different proctoring services, although there is commonality in the underlying AI technology at work.

Inspera Proctoring can record candidates via their webcam and will flag behavior it thinks potentially breaches exam conditions. This it does through implementing the use of AI in facial recognition and behavior analysis. For instance, students are required to take a picture of themselves using their webcam before the online assessment begins. This image can then be used to validate their identity. This allows a live proctor who sits in and invigilates a student or group of students via webcam to validate that the person taking the exam is who they say they are.

But this can also be used alongside AI behavior analysis to detect when the student’s face disappears from the view of the webcam. Inspera Proctoring would flag all of these occurrences with timestamps as part of the proctoring service. This is to allow a human to review the video either in a live proctoring or recorded exam proctoring environment to determine whether any rules were broken in these instances.

 

 Environmental Monitoring

AI also uses webcam feeds to assess a test-taker’s environment to determine whether there are any elements or materials that could be construed as suspicious or an indication of cheating.

For instance, closed-book exams would indicate that desks should be clear of notes and other sources of information, AI flags can indicate where a candidate has a book or phone present in their exam environment.

Environmental monitoring also includes noise detection. AI is capable of analyzing audio feeds, This can help to identify when there are others present in the room, offering the test-taker unauthorized assistance.

 

Data Privacy and AI Detection False Positives

We would be remiss if we failed to address concerns from both educators and students around AI proctorong.

Initial issues with proctoring seemed to center on privacy concerns. Candidates perceived, and some still do, that proctoring solutions were designed to spy on them and catch them cheating. One notable case is Ogletree v. Cleveland State University – where an Ohio judge ruled in favor of chemistry student Aaron Ogletree, who argued that the university’s virtual scan of his bedroom before an exam violated his rights under the Fourth Amendment. These constitutional rights are designed to protect US citizens against “unreasonable searches and seizures.” The case is currently being appealed and could set a precedent for future privacy laws and in turn, how data is used by online assessment software.

At Inspera we are dedicated to providing transparency through our proctoring services as well as how data is handled and shared. We have built in functionality to ensure users can maintain control over their data. Educators and administrators can set permissions to enable data sharing across institutions. This ethos is shared across our integrity tools, like in the similarity detection process of Inspera Originality.

However, educators can also grant students permission to delete their documents from this shared repository or remove the sources where similarities were detected. In this way, students are allowed control over their data, while maintaining academic standards and integrity.

That is also why we always emphasize the role that human judgment plays alongside AI when it comes to reviewing unusual behavior during exams. Our proctoring solution is not designed to detect cheating. Its purpose is to ensure exam integrity by alerting proctors of potential areas to review.

At no level should AI be used to determine whether a student has “cheated” during an exam. Obviously there are a variety of possibilities where AI might flag innocent behavior as a false positive. For example, a student leaving their workspace for a legitimate bathroom break would be flagged as they are detected as leaving their workspace, but a human review could determine whether this was cause for concern. That is why Inspera practices and encourages transparency with their educators and students. It is essential that students feel that they are not being “spied upon” in an effort to be “caught out.”

Instead, we urge educators to reinforce the benefits of online exam proctoring to students, including the increased accessibility these can offer test-takers. By upholding these values, AI can be an invaluable part of proctoring and the wider online assessment process.

 

Want to Know More About Inspera Proctoring?

If you’re interested in learning more about delivering online assessments with integrity then book a demo with us now and see how our proctoring services can integrate seamlessly with your requirements.

Book a demo

Written by Jo Bowden

January 29, 2024