Guide
The state of AI interview cheating.
A subscription industry now sells candidates the ability to cheat a live interview — invisibly. This is what changed, how candidates do it on and off the device, why the usual defenses miss it, and how to catch and deter it.
What changed
Remote interviews stopped being trustworthy.
For a monthly subscription that costs less than dinner, a candidate can now run a tool built for one purpose: to feed them expert answers during a live interview without anyone noticing. Some render answers on an invisible overlay; some listen through the browser; some replace the face on camera entirely. Many market themselves, openly, as “undetectable” on Zoom, Teams, and Meet.
The economics guarantee it keeps growing: cheap tools, a six-figure prize, and a new entrant almost every week. The result is measurable.
On the device
How candidates cheat — on the machine.
These are the common on-device vectors. Capifiq detects any unauthorized assistance running on the candidate's computer, not just a fixed list. Explore each:
Hidden AI assistants
Invisible overlays that feed answers in real time. Cluely, Interview Coder, and more.
Remote-control sessions
Someone off-camera viewing or driving the candidate's machine.
Virtual machines
A clean interview in a VM while tools run outside it.
Virtual cameras & deepfakes
A substituted, looped, or deepfaked webcam feed.
Browser-based assistants
“Interview copilots” that generate answers in a tab.
Display tampering
A clean shared screen hiding the real workspace.
The blind spot
Why the usual defenses miss it.
Screen sharing only transmits the surface the candidate shares — anything outside it, or on a second device, never appears.
Browser proctoring lives in the tab and can't see native desktop applications or the wider system.
Behavioral analysis infers from gaze and timing — indirect, and slower to adapt than the tools it chases.
Denylist scanners only catch catalogued tools; a rename or a new build slips past.
The pattern
- The tool runs where the defense can't see it
- Or it's renamed, private, or brand new
- Or the help never touches the device at all
The response
Catch what's on the device. Deter what isn't.
Interview integrity has two halves, and both are needed. Capifiq covers each.
Detect · on-device
A companion agent
Surfaces any unauthorized assistance on the candidate's machine — by what it does, not just its name — and seals it as timestamped evidence.
How detection works →Deter · off-device
Interviewing methodology
A phone, an earpiece, an accomplice — help no software can see. Deterred through how the interview is run.
How deterrence works →Go deeper
For your team, and versus the alternatives.
For technical hiring
Coding rounds are the number-one target. Keep them trustworthy.
For staffing agencies
Protect every placement and your reputation with sealed evidence.
For enterprise
Audit-ready integrity at scale, with no platform migration.
vs behavioral proctoring
A suspicion score, or sealed evidence you can act on.
vs lockdown & scanners
Restriction and denylists, or environment verification.
Security & trust
What is collected, how it's protected, and for how long.
Get started
See what your screen share can't.
Run Capifiq alongside the Zoom, Teams, or Meet calls you already do, and judge the evidence for yourself.