Can lockdown browsers stop real-time answer bots?

Not on their own. Lockdown browsers control the exam device, but the solver usually runs on a second device outside their reach. They are one layer, useful against casual cheating and essential for session integrity, but they do not see the phone on the desk.

Does AI text detection catch solver-assisted answers?

Unreliably. The test-taker transcribes or paraphrases the delivered answer, and short exam answers carry too little signal for detectors to judge. Timing and behavior analysis are more dependable than trying to detect the answer's origin.

Home / Blog

How do cheaters use a second device during proctored exams?

September 29, 2026 - ExamFence

Short answer: Second-device cheating happens when a test-taker keeps a phone, tablet, or second laptop just outside the webcam's view and uses it to look up answers, message a helper, or run AI tools during a proctored exam. It defeats most proctoring because the monitored device looks completely clean: no suspicious tabs, no unusual processes, nothing for lockdown software to flag. Detection has to look for indirect signals instead: gaze patterns that repeatedly leave the screen, audio anomalies like typing or notification sounds, network traffic from unknown devices on the same connection, and answer patterns that do not match the test-taker's demonstrated pace. The most effective countermeasure is environmental: requiring a 360-degree room scan and a clear desk policy before the exam starts.

Why the second device is the proctor's blind spot

Proctoring software monitors the device it runs on. Lockdown browsers control the tabs, the clipboard, and the running processes on that machine. But the test-taker's phone sitting on their lap is not running the lockdown browser, and no software on the exam laptop can see it. The entire security model assumes the monitored device is the only device, and that assumption fails the moment a second screen enters the room.

Cheating services have industrialized this. They sell playbooks that specify exactly where to place the phone relative to the camera, how to mute it, and what apps to use. Some setups go further: a helper on a video call who reads questions from the test-taker's screen share and dictates answers, all through an earbud hidden under hair. The exam device shows a model test-taker. The cheating happens in the room around it.

The signals that give it away

Since the device itself is invisible, detection works on behavior. Eye-tracking flags gaze that repeatedly drops below the screen or fixes to one side. Audio analysis catches what the camera cannot: keyboard sounds with no corresponding typing on the monitored device, notification pings, or the muffled cadence of someone listening to an earbud. These signals are individually weak and collectively strong.

Answer patterns add another layer. A test-taker who struggles with medium questions but nails the hardest ones, or whose answer timing shows long pauses followed by confident responses on items that should take working time, is showing the signature of external help. Network monitoring can sometimes catch the second device directly when it shares the test-taker's Wi-Fi, revealing an unrecognized device active during the exam window. No single signal proves cheating, but proctoring platforms score them together.

What institutions actually do about it

The highest-impact measure is also the lowest-tech: the environment check. A live 360-degree room scan with the webcam, showing the desk surface, the area under the desk, and the walls, catches most casual second-device setups before the exam begins. Clear-desk policies that require phones to be placed face-down across the room, verified on camera, remove the opportunity rather than trying to detect it.

For high-stakes exams, some programs add a second camera, usually the test-taker's phone itself, positioned to show the workspace from the side. This turns the most common cheating tool into a monitoring tool, which is elegant and effective. Others use audio baselines: recording ambient sound during the check-in and flagging deviations during the exam. The common thread is treating the room as part of the exam environment, not just the device.

Where detection is heading

The arms race continues on both sides. Cheaters adopt smaller earbuds, phones with silent haptic feedback, and AI tools that answer from a single glance at the screen. Detection responds with better gaze estimation, on-device audio classification that distinguishes typing on the exam laptop from typing nearby, and behavioral models trained on confirmed cheating sessions rather than rules of thumb.

The structural answer is exam design. Questions that require applying concepts to novel scenarios are harder to look up or dictate than factual recall, and they compress the value of a second device. Time pressure tuned to the honest test-taker leaves less room for the lookup-and-wait cycle. Technology will keep improving on both sides, but exams that are hard to cheat on beat surveillance that tries to catch cheating after the fact.

Can lockdown browsers detect a phone nearby?

No. A lockdown browser controls the device it is installed on and has no visibility into other devices in the room. Detecting a second device requires camera, microphone, or network signals, which is why environment checks and behavioral analysis matter more than device lockdown for this threat.

Is requiring a second camera too invasive?

It is the most effective technical countermeasure, but it raises real privacy concerns that institutions have to weigh. Many programs reserve it for high-stakes professional exams and use room scans plus behavioral flagging for lower-stakes assessments. The right level depends on what the credential is worth.

Do cheaters get caught after the fact?

Sometimes. Statistical analysis of answer patterns can flag sessions for review, and institutions do invalidate scores when the evidence is strong. But post-exam detection is slower and more contested than prevention, which is why the emphasis has shifted to environment controls before the exam starts.