Skip to main content

Anna Anti-Fraud and Interview Integrity

Understand Anna settings that help discourage cheating and support human review.

Overview

Anna (AI Recruiter) includes optional settings that can help discourage or flag suspicious interview behavior.

These tools are designed to support human review. They should not be used as the sole basis for rejecting candidates.


Available Interview Integrity Settings

Most of these settings are configured per job opening. Unnatural pause detection is enabled for your account by Fountain Support.

Setting

Description

How to enable

Fullscreen mode

Forces the interview to be conducted in a full-screen browser window. Candidates who exit fullscreen are flagged.

Checkbox on the job opening

Record video

Records candidate video during the interview

Checkbox on the job opening. Candidate must provide webcam access and use a desktop computer

Unnatural pause detection

Flags repeated extended pauses as a possible cheating risk

Optional. Contact Fountain Support to enable it for your account

Restrict interview link access

Only allows pre-registered candidates to interview

Checkbox on the job opening. Requires uploading candidates in advance

AI interviewer instructions

Sets expectations about interview behavior

Text field on the job opening. Deterrent only


Fullscreen Mode

Fullscreen mode helps discourage candidates from switching tabs or using other windows during the interview. Candidates who exit fullscreen mode are flagged for review.

Limitations:

  • Applies to browser-based interviews, not phone interviews

  • Does not prevent all forms of outside assistance

  • A fullscreen exit is an integrity signal, not proof of cheating


Record Video

Video recording allows reviewers to see the candidate during the interview.

Requirements:

  • The candidate must use a desktop computer

  • The candidate must provide webcam access

  • The candidate should be informed before recording begins

Use recordings to support review, especially when there are concerns about identity, interruptions, or suspicious behavior.


Unnatural Pause Detection

Unnatural pause detection flags candidates with a possible cheating risk when they repeatedly take an unusually long time to respond. Long gaps can indicate a candidate is looking up or generating answers rather than speaking from experience.

This setting is optional and is not enabled by default. Contact Fountain Support to turn it on for your account.

A pause flag does not prove cheating. Candidates pause for many ordinary reasons, including nerves, background noise, a poor connection, or simply thinking. Treat the flag as a prompt to review the transcript and recording more closely, not as a reason to reject.


Restrict Interview Link Access

By default, anyone with the interview link can start an interview. Enabling Restrict interview link access limits interviews to candidates you have uploaded in advance, and blocks unrecognized candidates.

This is the strongest control against someone other than your intended candidate taking an interview. It also adds an upload step to your process, so weigh it against your volume and workflow before enabling it.


Deterrence Through Instructions

Anna cannot guarantee detection or prevention of all AI assistant usage, such as ChatGPT Voice, Siri, or other tools.

Recommended approach:

  • Use fullscreen mode and video recording when appropriate.

  • Ask Fountain Support about enabling unnatural pause detection.

  • Add deterrence language to AI interviewer instructions and candidate messaging.

  • Review transcripts and recordings for suspicious patterns.

  • Use human follow-up when needed.

Example deterrence language:

"Please complete this interview using your own answers. Use of AI voice assistants or other outside tools during the interview may result in additional review."


Review the Transcript and Recording

Every completed interview includes a full transcript and a recording. The transcript view supports playback speed control and translation, so a reviewer can work through an interview quickly or in another language.

Signals worth a closer look:

  • Answers that read as scripted or noticeably unlike the candidate's other responses

  • Repeated long gaps before answers

  • A fullscreen exit flag

  • Audio that suggests another person is present or prompting


Respond to Flagged Behavior

  1. Review the transcript.

  2. Review the recording, if available.

  3. Compare responses against role qualifications.

  4. Consider a human follow-up interview.

  5. Document the review according to your organization's policy.

Do not reject a candidate automatically based only on an integrity flag.

Did this answer your question?