How we prevent AI cheating for remote proctoring tests

Who can blame them? Gen AI explains complex topics, improves writing, brainstorms ideas, and more. All within an instant.

When it’s time to take an assessment, however, the accuracy and speed of generative AI makes it a potent cheating method. So it must be blocked – which is exactly what Janison does for its remotely-proctored tests.

In this article, we explain how our remote proctoring solution can block gen AI for your students, helping to preserve your academic integrity.

How we prevent AI cheating for remote proctoring tests

Our remote proctoring tests are delivered through a software app, which students (or your IT team) download to their devices.

During tests, the app monitors the student’s operating system to watch for gen AI programs being launched. If they try to open ChatGPT or another AI tool, our app immediately blocks it and flags the act as a violation, which one of your invigilators can review. The student can also be warned on the spot or even kicked out entirely, depending on how you want to enforce the rules.

In addition to this software app, our online proctoring solution includes a web browser extension, which students must also install prior to starting their tests. If they try to access gen AI through Chrome, Firefox, or another web browser, it’s detected and instantly blocked.

Together, the app and locked-down browser extension block ChatGPT, Gemini, Claude, Copilot, Meta AI, and every other popular generative AI tool, removing this powerful cheating method for students.

What if students try to access gen AI on another device?

As students take their tests, our remote proctoring software uses their webcam to monitor and record them, and it automatically flags whenever they use their phone or another device.

This creates a timestamp in the video recording, which a staff member reviews once the test is over. They can determine whether the student really did use another device and for how long, and refer to their QA guidelines to decide the penalty.

If you’d like to learn more about creating rigorous QA guidelines to review these kinds of flagged violations, check out section 3 of this blog.

Can students bypass the gen AI restrictions?

The application and browser extension are a requirement for the remote proctoring test, and the software is locked. That means students can’t change its settings and bypass the AI restrictions.

Comprehensive cheating prevention

Generative AI is just one possible cheating method for remotely proctored tests. Our software solution blocks every common technique, including:

• The internet, including specific websites
• Any word-based apps that can store notes
• Virtual machines (e.g. another instance of the Mac operating system running on the same machine)

In addition, by monitoring through the students’ webcam, the system accurately detects common cheating methods outside of the system like proxy-test taking, looking at physical notes, messaging apps on mobile phones, and more.

When used in combination with the technical blocks imposed by the software, our remote proctoring solution reduces or completely blocks the vast majority of cheating. It allows institutions and companies to retain their academic integrity with online proctoring – including defending against generative AI.