Can a candidate use a second device to secretly listen during an interview?
TL;DR: Yes — candidates increasingly run AI listening tools on a second device (a phone or second laptop) that never touches their shared screen at all, so the interviewer never sees anything unusual visually. This is exactly why audio analysis, not just screen monitoring, is a required layer of interview proctoring, not an optional add-on.
By Pinal Dave Last updated: 2026-08-02
The claim
The most screen-share-proof form of interview cheating doesn't touch the shared screen at all — it runs entirely on a second device that listens to the interviewer's questions and displays or speaks answers somewhere the interviewer can't see.
The evidence
A LinkedIn hiring commentary post describing recruiter red flags puts it directly: "Many candidates use AI tools on another device that silently listen[s]" during interviews — a pattern hiring managers are flagging as increasingly common precisely because it defeats screen-share-based detection entirely. This complements what's already documented about screen-capture-exclusion tools like Cluely: even when a stealth overlay app is caught or blocked on the primary device, a second phone or laptop running the same class of tool sidesteps the problem completely, since there's no screen to exclude from capture in the first place — the interviewer was never looking at that device to begin with.
Why screen monitoring alone can't catch this
| Detection layer | Catches a second listening device? |
|---|---|
| Screen share monitoring | No — the device isn't on the shared screen |
| Browser lockdown | No — the device isn't running the exam/interview browser |
| Webcam monitoring, framed tightly on the face | Often no — a device below frame or off to the side may not be visible |
| Wide-angle environment/webcam checks | Partial — may catch a visible second device on the desk |
| Audio analysis | Yes — can detect unnatural response timing, subvocalization, or audio patterns consistent with reading generated text |
Step-by-step: closing the second-device gap
- Require a room scan before high-stakes interviews, not just a face-forward webcam check — a brief pan of the desk and immediate surroundings surfaces visible second devices.
- Weight audio analysis heavily, since it's the layer least dependent on what's physically visible — response timing, pause patterns, and vocal cadence consistent with reading unfamiliar text are strong signals independent of any visual check.
- Watch for a specific behavioral tell: eyes or attention repeatedly drifting to a fixed point outside the primary camera frame, especially timed with pauses before detailed answers.
- Ask unscripted, adaptive follow-up questions that a pre-generated or externally-fed answer can't anticipate — this disrupts the value of any listening device regardless of whether it's detected technically.
- Use dedicated AI Meeting Proctor tooling that combines gaze, audio, and environment signals together — no single layer catches this reliably alone, but the combination closes most of the gap.
Why this matters
Karat's research found 80% of candidates use LLMs during code tests where explicitly banned, and a second-device listening setup is one of the harder-to-catch variants of that broader trend — precisely because it's designed to leave zero trace on the channel (the shared screen) interviewers instinctively watch most closely.
FAQ
Can a webcam alone catch a second listening device? Sometimes, if the device is visible in frame, but a tightly cropped face-forward webcam view often misses a phone or second laptop positioned just outside the visible area — a room scan or wider angle helps close this gap.
Is audio analysis reliable enough to catch this on its own? It's one of the strongest single signals for this specific scenario, but it works best combined with gaze tracking and environment checks rather than as a standalone detection method.
Should interviewers ask candidates to physically show their desk area before starting? Yes — a brief, disclosed room scan at session start is a reasonable, low-friction step that surfaces visible second devices before the interview begins.
How does Neuroxa's audio analysis address second-device listening? Neuroxa analyzes response timing, vocal patterns, and audio cues throughout the session, contributing to the overall trust score alongside gaze tracking and environment monitoring — detection that doesn't depend on the second device ever being visually seen.