Candidates may attempt to cheat using mobile phones, external devices, AI tools, screen sharing, hidden notes, virtual machines, online searches, or by receiving help from someone nearby. Cheating can be reduced by continuous screen and webcam proctoring, identity verification, and lockdown browser to make unauthorized activity easier to detect and limit opportunities for external help.
The candidate shares their username/password with another person. The other person logs in and completes the exam on their behalf. This is often done by a friend or a paid “exam taker,” also known as a proxy test taker.
ID verification requires the candidate to prove their identity using both an official document and their live facial appearance.
During the verification process, the candidate must provide:
The system then uses AI-based facial matching to compare the face on the ID with the live face in front of the camera. If the live face does not match the ID photo, the system immediately flags the attempt or blocks entry to the exam.
Proctoring continuously monitors the candidate’s identity, behavior, surroundings, and computer activity throughout the entire exam. It makes it extremely difficult for someone other than the registered candidate to take the test or assist them from off-camera.
Proctors watch the candidate in real time through the webcam.
This prevents impersonation because:
Candidates place a phone on their lap, just below the webcam’s view, and repeatedly look down to search answers on Google, ChatGPT, or notes. They quietly use WhatsApp, Telegram, or similar apps to message someone who sends answers back in real time, usually with the phone held low or under the desk. They quickly take pictures of questions with a phone and send them to a friend or tutor who replies with the correct answers.
They place the phone directly behind the laptop screen or slightly to the side, then lean back or tilt their head slightly to read answers without the camera noticing. A second device, such as another laptop or a tablet, is placed outside the webcam frame, usually to the right or left of the main screen, and is used to look up answers or run AI tools.
Modern platforms combine AI proctoring, live human proctoring, multi-camera monitoring, and screen proctoring to eliminate blind spots and detect suspicious behavior instantly.
Proctoring works by continuously monitoring the candidate’s:
AI-proctoring analyzes the video feed to detect patterns commonly associated with device usage, such as:
Live human proctors reinforce this by visually identifying behaviors AI may flag, including:
They can issue warnings, ask for a repositioning of the camera, request a room scan, or stop the exam if device usage is confirmed.
Multi-camera proctoring removes typical cheating blind spots. The laptop webcam shows the candidate’s face and upper body, while a second camera (often the candidate’s phone) captures their hands, lap, desk surface, and the sides of the laptop. This prevents common device-based cheating methods such as:
Candidates keep the exam open in one tab, then open a new tab to search answers on Google, ChatGPT, YouTube, or question-bank websites. They switch back and forth quickly, hoping the proctor or system won’t detect the tab changes.
Instead of using regular tabs, some candidates open incognito or private windows. They believe private windows are less detectable and leave no browsing history, making it harder for a simple monitoring setup to catch them. They keep the exam visible on one window while searching in an incognito window next to it or behind it.
A browser lockdown completely restricts what a candidate can do on their computer during the exam. It forces the exam to run in a controlled, isolated, full-screen environment, and prevents the candidate from accessing any external websites, applications, or tools.
When the lockdown browser is active, the system automatically disables or blocks:
Real-time screen proctoring gives the proctor or AI a live view of the candidate’s entire computer screen, making every attempt to access Google, ChatGPT, or any other online resource immediately visible.
Because the screen is continuously monitored, candidates cannot secretly open new tabs, switch browsers, or hide search windows behind the exam. Every on-screen action is captured the moment it happens.
If a candidate tries to:
The proctor or AI system instantly sees the new tab or window pop up on the screen. No browsing activity can be hidden.
Candidates often misuse AI by taking the question they see on the exam screen and sending it to an AI chatbot for an instant solution. After reading the exam question, they either manually type it or copy–paste the text into an AI tool. This can happen in several ways:
Once the question is entered into the AI chatbot, the model immediately generates the solution, often faster and more accurately than manual searching. The candidate then returns to the exam window and uses the AI-generated response.
By combining AI monitoring and human oversight, proctoring exposes every attempt to access or use AI, whether on the exam device or an external device like a phone.
Proctoring blocks AI cheating in two major ways:
AI and human proctors continuously watch the candidate’s screen and webcam feed. This exposes any attempt to use AI directly on the exam device, including:
Even brief flashes of an AI window are visible to the proctor.
Proctoring continuously monitors the candidate’s face, upper body, eye movement, and behavior, making it difficult for them to use AI tools on external devices without revealing visible cues.
Proctors can detect suspicious physical behavior linked to AI use, such as:
Candidates attempting to use a phone on their lap, a tablet or device placed beside the exam laptop, AI responses displayed just outside the camera’s view, inevitably reveal non-natural viewing angles, repeated glances, or physical motion that is inconsistent with focused exam behavior.
These behavioral cues become obvious indicators of external AI use, allowing human or AI proctors to flag, warn, or terminate the session even with just a single camera.
Some candidates launch or join a hidden video call where they secretly share their screen with a friend, tutor, or hired helper. The helper watches the exam questions appear, researches or solves them, and feeds back the answers through text or voice. Candidates often place the call window behind the exam window or on another desktop.
Candidates sometimes install remote-access software before the exam.
When the exam starts, the helper can see the exam questions in real time and guide the candidate by giving answers. The candidate remains on camera to look like they’re taking the test themselves while someone else essentially “co-pilots” from afar.
Instead of using apps, some candidates mirror their laptop screen to a TV, another computer, a tablet, or a phone placed elsewhere in the room. This allows another person to sit off-camera and watch the entire exam.
A lockdown browser it restricts the candidate’s device to a single, secure testing environment. When the lockdown browser is active, candidates cannot launch, access, or interact with any software or feature that could broadcast their screen to someone else.
A lockdown browser blocks every major application that could be used to share the candidate’s screen with someone else. It prevents launching or running tools such as:
Even if these applications are already open in the background, the lockdown browser automatically forces them to close when the exam starts, preventing anyone else from viewing, controlling, or accessing the candidate’s screen.
A helper can simply stay in the same room, sit or stand wherever they want, and support the candidate throughout the exam. They may sit next to the desk, behind the candidate, near the door, or anywhere else in the room where they can comfortably see or hear the questions.
The helper can read the questions directly from the candidate’s screen or wait for the candidate to read them out loud.
From there, the helper provides the answers openly by speaking normally, discussing the question, explaining the reasoning, or giving instructions without restriction. If needed, they can also use books, notes, or online tools to look up the information and deliver the correct response.
By monitoring the candidate’s environment, behavior, and audio in real time, proctoring makes it extremely difficult for another person to remain in the room without being detected.
Proctoring continuously analyzes the candidate’s eye movement, body posture, and surrounding audio to detect signs of in-room assistance. Human and AI proctors look for patterns such as:
These visual and audio indicators together make it easy for proctors to spot when a candidate is receiving help from someone in the room.
Candidates often rely on physical or digital notes placed outside the monitored area to access information during an online exam.
Proctoring makes it extremely difficult for candidates to use hidden notes by monitoring their environment, body language, and behavior throughout the exam.
Effective prevention includes:
Candidates run the exam inside a virtual machine and use the real operating system to browse, search, or run unauthorized tools.
Common intentions include:
Candidates use macOS Mission Control or Windows Task View to create:
Screen proctoring prevents candidates from using virtual machines or multiple desktops by showing exactly what happens on their device in real time. Proctors and AI can immediately spot:
Because every on-screen action is visible, any attempt to access an unmonitored environment becomes clear and is flagged instantly. This makes it extremely difficult for candidates to hide secondary desktops, run VMs, or operate unauthorized tools during the exam.
Candidates often use chat applications on another device to send questions and receive immediate answers. Some candidates join group chats where multiple people help simultaneously. These groups discuss questions, cross-check answers, solve problems collectively, share reasoning or explanations. The candidate follows the group’s suggestions and enters the answer in the exam.
Linear-on-the-fly testing prevents candidates from collaborating with others by giving every test-taker a unique, personalized version of the exam. Because no two candidates receive the same set of questions in the same order, coordinating answers with friends, study groups, or online helpers becomes extremely difficult.
LOFT reduces collaboration by:
With LOFT, even if candidates try to collaborate via chat groups, messaging apps, or shared notes, their exams will not match, and the answers provided by others will often not apply. This breaks the efficiency of real-time group support and makes collaboration highly unreliable.
Proctoring makes it extremely difficult for candidates to collaborate with others during an online exam by monitoring their behavior, audio, surroundings, and interaction patterns. Even if a candidate tries to communicate through a phone, another device, or directly from their computer, proctoring detects the cues that reveal external help.
Proctoring detects collaboration through:
These combined visual, audio, and on-screen indicators make collaboration easy to identify, allowing proctors to flag or stop the exam immediately.
Yes, modern online exams can detect cheating using a combination of webcam monitoring, screen proctoring, audio detection, browser lockdown, and AI-based behavior analysis. These systems can identify actions like tab switching, screen sharing, use of virtual machines, unusual movements, background voices, hidden devices, and attempts to collaborate with others. When combined, these tools make it possible to detect and often prevent most cheating attempts during an online exam.
Yes, online exam systems can detect when a candidate opens another tab, switches browsers, or navigates away from the test window. Screen proctoring and browser lockdown tools monitor tab changes and flag any attempt to access external websites or resources. Some platforms also generate automatic warnings or restrict the exam if repeated tab switching occurs.
Yes, it is possible, but it is extremely difficult. Modern online proctoring systems use multiple security layers, including webcam monitoring, screen proctoring, audio detection, browser lockdown, AI-driven behavior analysis, and even multi-camera setups. These measures make most cheating attempts easy to detect, from using external devices to collaborating with others or switching screens.
While no system can guarantee 100% prevention, a well-designed online proctored exam significantly reduces the chances of cheating and ensures a high level of exam integrity.
Yes, it is possible, but most attempts are blocked or detected. A lockdown browser is designed to block the most common cheating methods by preventing new tabs, applications, keyboard shortcuts, screen sharing, and external navigation.
However, a lockdown browser alone cannot prevent every type of cheating. Without additional measures like proctoring, candidates could still attempt methods that happen outside the device, such as using hidden notes, external devices, or getting help from someone off-screen.
Yes, it is possible, but the opportunities are extremely limited. Two-camera proctoring gives a full view of both the test-taker and their surroundings, making it much harder to use hidden notes, external devices, or get help from someone off-screen. The front webcam monitors facial movements and on-screen behavior, while the secondary phone camera captures the desk, hands, and room environment.
With both angles monitored (often combined with audio detection and AI analysis), most cheating attempts become immediately visible. While no system can guarantee 100% prevention, two-camera monitoring significantly reduces the chances of cheating and provides a much higher level of exam security.
