Are Your Assessments Truly Secure?
❌ Do you already have an assessment vendor, yet candidates still manage to cheat?
❌ Do your campus hiring tests let unqualified candidates slip through?
❌ Have your test questions been leaked, compromising hiring decisions?
❌ Do candidates use AI-powered tools like ChatGPT to find answers?
Traditional AI proctoring solutions rely on rule-based detection methods that fail to adapt to evolving cheating techniques, such as AI-generated voice assistance or screen-sharing hacks. They often generate false positives or miss subtle cues like whispering or hidden devices. Sherlock, on the other hand, uses multi-modal intelligence—combining vision, hearing, and cognitive analysis—to detect even the most sophisticated cheating methods in real time. By offering context-aware alerts and adaptive monitoring, Sherlock ensures a tamper-proof testing environment that traditional systems cannot achieve.
Sherlock by WeCP is an advanced AI-powered proctoring and test integrity system designed to detect and prevent cheating in remote assessments. It ensures a fair evaluation environment by using computer vision, audio analysis, and cognitive detection mechanisms. This document explains how Sherlock works and what makes it a powerful solution for test integrity.
1. How Sherlock Works
Sherlock actively monitors a candidate's activity during an assessment, analyzing multiple data points to detect suspicious behavior. It operates using three core pillars:
Vision: Detects external devices such as mobile phones, additional screens, multiple monitors, and unauthorized individuals.
Hearing: Identifies human conversation, spoken keywords related to the test, and AI-generated voices.
Cognition: Analyzes behavioral patterns such as abnormal typing speed, excessive tab switching, and unusual response patterns.
Sherlock provides real-time alerts to administrators about detected anomalies, allowing test supervisors to take immediate action.
Key Monitoring Features
Device Detection: Identifies the presence of multiple keyboards, mice, screens, and virtual machines.
Screen & Clipboard Monitoring: Prevents unauthorized screen sharing, copy-pasting, and screenshots.
Audio Proctoring: Recognizes spoken words and flags suspicious phrases like "ChatGPT" or whispered answers.
Environmental Awareness: Detects multiple faces or people within the test environment.
Activity Logging: Maintains an audit trail of all activities, timestamped for later review.
Example Sherlock Alerts:
{sherlock_says: "A mobile device is spotted at 20th minute of the test for this user"} {sherlock_says: "Heard someone speaking the word 'coding question' at 35th minute"}
2. Why Sherlock is Essential for Test Integrity
Sherlock ensures a tamper-proof testing environment by:
Eliminating unauthorized aid from external devices or individuals.
Providing automated, unbiased monitoring to detect and deter cheating.
Enabling scalable proctoring for high-volume testing environments.
Reducing the need for human proctors while maintaining test security.
Sherlock is designed for: ✅ Pre-employment assessments
– ensuring fair hiring decisions. ✅ Certification exams – upholding credibility and preventing fraud. ✅ Academic testing – maintaining integrity in remote learning assessments.
3. How Sherlock Detects and Reports Violations
Real-Time Monitoring & Alerts
Sherlock continuously scans the test-taker's environment and actions, instantly flagging violations. These alerts are sent to test administrators for review.
Violation Categories
Low-Risk Violations – Minor anomalies like a candidate glancing away briefly.
Medium-Risk Violations – Unusual sound detection or short screen changes.
High-Risk Violations – Presence of multiple faces, mobile phone detection, or spoken answers.
Each violation is logged and time-stamped to provide a clear history of candidate behavior during the test.
Example Logged Events:
10:05 AM: Candidate switched tabs. - 10:20 AM: External monitor detected. - 10:35 AM: Candidate whispered a possible answer.
4. Security & Compliance Considerations
Sherlock is built to meet stringent security and privacy requirements:
Data Encryption: All test data is encrypted to prevent unauthorized access.
User Privacy: Sherlock follows GDPR, CPRA, and global data protection guidelines.
Access Control: Only authorized administrators can review flagged violations.
5. Summary & Next Steps
Sherlock by WeCP is a next-generation AI-powered test integrity solution that ensures fair and reliable assessments. By combining real-time monitoring, anomaly detection, and automated reporting, it provides organizations with a scalable and effective way to uphold test security.
Next Steps:
Set up Sherlock in your assessment environment.
Configure alerts and detection sensitivity based on test requirements.
Monitor flagged events and take necessary actions.
Continuously update test policies based on Sherlock insights.
See Sherlock in Action:
This short video walks you through how Sherlock detects AI-assisted cheating using advanced vision, hearing, and behavior analysis during remote interviews.
For further details, visit the Sherlock Knowledge Base or contact WeCP’s support team.