Important: These indicators are completely independent of the AI interview score. They do not affect a candidate's score, ranking, or recommendation in any way.
Security flags exist to support human review — not to make decisions automatically. When a flag is triggered, recruiters should review the associated evidence (video, audio, or activity logs) before drawing conclusions. Depending on your organization's hiring policies, confirmed instances of interview fraud or policy violations may result in candidate disqualification.
1. Third-Party Voice Detection
What it checks: Whether additional voices are present during the interview.
This includes coaching, prompting, or conversations occurring while the candidate is responding.
Status | What it means |
| Only the candidate's voice detected. |
| Possible background voices detected. |
| Third-party voice detected during one or more responses. |
2. Eye Gaze Analysis
What it checks: Whether the candidate generally looks towards the camera while responding.
This is an attention signal. Looking away from the camera may have been caused by many factors, such as thinking or reading notes. This indicator should always be interpreted together with other security signals.
Status | What it means |
| The candidate looked towards the camera for 75% or more of the response. |
| The candidate looked towards the camera for 50–74% of the response |
| The candidate looked towards the camera for less than 50% of the response. |
3. AI Content Detection
What it checks: Whether responses show characteristics commonly associated with AI-generated or AI-assisted content.
Status | What it means |
| No significant indicators of AI-generated content. |
| Some indication of AI-generated content. |
| Strong indicators that one or more responses may have been AI-generated or substantially assisted by AI. |
4. Lip Sync Detection
What it checks: Whether the candidate's lip movements are synchronized with the recorded audio.
This helps identify manipulated or synthetic (deepfake) video responses.
Status | What it means |
| Lip movements are consistent with the recorded audio. |
| Minor synchronization inconsistencies detected. |
| Significant lip-sync inconsistencies detected. |
5. ID Verification
What it checks: Whether the person completing the interview matches the identity document submitted.
When a candidate uploads a government-issued ID or other identity document, the system compares the face on the ID with the face in the video interview and generates an identity match score.
Status | What it means |
| Identity match score is 70% or higher. |
| Identity match score is below 70% — the submitted ID may not match the interview participant. |
6. Voice Detection
What it checks: Whether the candidate's responses contain clear spoken audio. This helps identify responses with missing audio, muted microphones, or extremely limited verbal answers.
Status | What it means |
| Clear voice detected in all responses. |
| One or more responses contain very limited speech (fewer than 10 spoken words). |
| No voice detected in one or more responses. |
7. Browser Tab Switching
What it checks: Whether the candidate left the interview window while recording a response.
Frequent tab switching may indicate that external resources were accessed during the interview. This signal should be interpreted together with other security indicators.
Status | What it means |
| No browser tab switching detected. |
| Limited browser tab switching detected. |
| Repeated browser tab switching detected. |
8. Copy & Paste Detection
What it checks: Whether text was pasted into text-based interview responses.
This helps identify answers that may have been prepared using external sources rather than written in real time.
Status | What it means |
| No copy-and-paste activity detected. |
| Limited copy-and-paste activity detected. |
| Significant copy-and-paste activity detected. |
How to Use These Signals
Review, don't reject. A flag is a prompt for human review, never an automatic disqualifier. Environmental factors — background noise, poor lighting, camera placement — can trigger false positives.
Check the evidence. Each flag links to the specific response or moment that triggered it. Watch or listen before making a judgment.
Look for patterns. A single yellow flag is rarely conclusive. Multiple red flags across different checks warrant closer scrutiny.
Follow your policy. Disqualification decisions should follow your organization's documented hiring and integrity policies.
Interviewer.AI — Smart Video Interviews, Trusted Hiring Decisions