Understanding AI-Based Proctoring System:-
1. Definition and Features:
An AI-Based Proctoring System is employed by the AI to disprove candidates in online exams. All activities related to identification of human candidates are on autopilot-from face detection, voice detection, browser tracking to even certifying the test-taker can complete a test in a relatively secure environment. Where traditional tools support more static monitoring, the modern ones offer more dynamic and real-time adaptability.
2. Types of AI-Based Proctoring:
- Live Proctoring: The human invigilator watches the proctoring process in real-time.
- Recorded Proctoring: The whole session is recorded and reviewed at a later time.
- Automated Proctoring: Fully AI-run with real-time flagging – truly scalable over and above the traditional platforms of Proctor U or Mercer | Mettl in a high volume assessment scenario.
What is Flagging in an AI-Based Proctoring System?
1. Meaning of Flagging:
Flagging is a term used for the detection and tagging of suspicious activities occurring during an exam. These flags are mostly ranked according to their significance.
2. Types of Flags:
- Low-Level Flag: Background noise or instance of distraction.
- Medium-Level Flag: Continual look-away.
- High-Level Flag: Another person is present, switching tabs, etc.
3. Examples of Flagged Behaviors:
- Whispering or unusually loud background noise.
- Face not detected for a long period.
- Using mobile phones or external devices – something often missed in the old generation systems like Mercer | Mettl.
Think Proctor Customization:-
1. About Think Proctor:
Think Proctor is the advanced AI-based proctoring system of Think Exam, providing flexible and intuitive customization tools to an examination administrator. Such systems give less flexibility to the candidates, while Think Proctor enables real-time adjustments.
2. Flexibility from Think Exam’s Proctoring Software:
Think Proctor’s AI-based proctoring system enables turning on and off monitoring features as required by the test being exercised and its importance – something less fluid in legacy such as Mercer | Mettl.
3. Setting Changes During Exams and Admin Oversight:
Administrators can change settings even while an exam is in progress to enable dynamic control over the proctoring activity, a level of dynamism not common in systems like Proctor U.
Steps to Customize Flagging Criteria in Think Proctor:-
Step 1: Login to Admin Dashboard:
Use your administrator credentials to enter the Think Proctor dashboard.
Step 2: Head Towards Proctoring Settings:
Head on to the settings tab and click on the “AI-based proctoring system configuration” option.
Step 3: Select From AI Flagging Options:
Choose the flag categories you want such as facial recognition, noise detection, and tab-switching.
Step 4: Set Threshold for Suspicious Activity:
Modify the sensitivity level and decide how many violations are allowed before a warning is issued or the test ends.
Flagging Customization Major Criteria in AI-Based Proctoring System:-
1. Voice Detection Sensitivity:
Admin can adjust the sensitivity of the AI-based proctoring system towards ambient sounds and the speech of the candidates.
2. Eye Movement and Face Detection:
Customize how long a candidate can look away or whether his/her facial presence has to be constant.
3. Browser Activity Monitoring:
Enable or disable tab-switch alerts and clipboard access restrictions.
AI Customizations for Different Types of Exams:-
1. Customizations for High-Stakes Exams:
Strict monitoring should be implemented, allowing no tolerance for tab switches or missing face.
2. Settings for Skill-Based or Open-Book Tests:
Fairly loose restrictions to allow reference materials while flagging others – Unlike Mercer | Mettl, Think Proctor is considered by many to be a better platform for this.
3. Remote Hiring Test Customization:
Adjust voice and face tracking, and allow use for IDEs or workspaces.
Incorporating Human Review With AI Flags:-
1. Human Proctors’ Roles in Verification:
Even the smartest AI-based proctoring system benefits from a healthy second human opinion in those edge cases.
2. Merging AI Flags and Manual Review:
Think Proctor provides a dual view—AI-based proctoring system summary plus session replay—for human cross verification.
3. Making Final Decisions:
Admins can come to a final decision after reviewing footage and AI reports generated by the flags.
Best Practices in Customizing:-
1. Always Run Test Simulations:
Pilot tests predict the practicability of your flagging rules in real-time.
2. Get Feedback from Proctors and Candidates:
It’s pertinent to set arrangements which will, in practice, work better due to real-time information relating to the flag implementation from proctors and candidates alike.
3. Continue Reviewing/Updating Set Rules:
Periodic audits need to be done to keep flagging rules aligned with the changes in test patterns.
Advantages in Enterprise And Education:-
1. Scalability for Universities and Corporates:
The Think Proctor’s AI-based proctoring system offers smooth scalability to thousands of concurrent users without delay.
2. Integrations with LMS Seamless:
It fits well into your current system through APIs and plug and play models.
3. Powerful Reporting and Analysis:
Look for analytical insights about candidate behavior and system performance.
Conclusion:-
Customizing flagging criteria in an AI-based proctoring system like Think Proctor is not merely a technical necessity; it’s also a strategic way because it enables institutions to ensure exam integrity while allowing certain testing conditions and formats. Whether it’s a certification exam or a mass hiring drive, well-tuned proctoring dissolves ambiguity and builds trust.
FAQs:-
1. What is Think Proctor?
- Think Proctor is an AI-based proctoring system by Think Exam with customizable options for secure and seamless online assessments.
2. How do I change flagging settings in Think Proctor?
- Login to the admin dashboard and go to proctoring settings under which you can adjust thresholds for different AI-based proctoring system detection features.
3. Can human proctoring be combined with Think Proctor’s AI?
- Yes, Think Proctor supports hybrid models in which human proctors review AI-generated flags.
4. Is Think Proctor good for open-book exams?
- Definitely! It allows you to relax certain flags and monitor behaviors indicative of other types of cheating.
5. Is customization affecting the performance or system load?
- Customization does not hamper the performance or put a load on the system for the Think Proctor’s AI-based proctoring system is designed with large-scale customizable configurations in mind.
