AI Image Proctoring in Interviews

AI Image Proctoring helps verify a candidate’s identity and continuous presence throughout an AI-led interview. It ensures interview integrity by capturing a baseline selfie at the start and periodic webcam snapshots during the session.

The system automatically detects and flags integrity risks such as:

• No face detected

• Multiple faces detected

• Unrecognized face detected

All violations are summarized in a Proctoring Report, highlighting overall risk levels as Green, Yellow, or Red.

Key Assumptions

• Image proctoring covers webcam monitoring only (no screen or keyboard tracking).

• Proctoring runs in parallel with the interview without affecting scoring.

• Candidates must grant camera access and provide consent before starting.

• The system automatically retries if network issues occur.

• All images and data comply with your organization’s data retention and privacy policies.

Configuring Image Proctoring

1. Enable Image Proctoring (Interview Creation Platform)

Go to recruitment.imocha.io → Interview Creation screen.

As an interview creator, you can enable or disable Image Proctoring when creating an interview invite and define default policies that apply to all interviews.

Settings available:

• Image Proctoring: Yes / No (Default: No)

• Default Violation Limit: Number input (Min: 0, Max: 20)

• Capture Interval: Select interval (e.g., every 20s–40s)

• Consent Text: Customizable rich text that appears to candidates before starting the interview

After configuring, click Save to apply these settings for all interviews created.

2. Per-Interview Proctoring Controls

When creating or editing an individual interview, you can override default settings.

Options include:

• Image Proctoring: Yes / No toggle (with tooltip explanation)

• Violation Limit: Enter number of allowed violations before auto-termination (enabled only if Proctoring = Yes)

• Capture Interval: Select snapshot frequency (enabled only if Proctoring = Yes)

These settings ensure flexibility for different roles or interview types.

3. Violation Detection Rules

The system flags violations based on captured images:

Auto-Termination Rule:
If total violations exceed the Violation Limit, the interview automatically ends.

Violation Count Logic:
Three consecutive images with violations count as one violation.