3 minute read  |  The deepfake candidate problem your company isn’t ready for

A hiring manager we recently spoke with described an interview that didn’t quite feel right.

The candidate’s video was slightly off. The lip movements didn’t perfectly match the audio. Answers came through just a beat too late. When pressed with a quick, unexpected question, the candidate paused awkwardly, then gave a polished response that didn’t quite address what was asked.

The hiring manager had the unmistakable feeling that the person on screen was not the person they were actually hearing.

That feeling is now being validated by a growing body of reports. Fake candidates are showing up to real interviews. Sometimes using AI-generated faces. Sometimes with someone else answering off-screen. Sometimes both. And most companies are not ready for it.

How this is actually happening

Three patterns keep showing up in reports from hiring teams in 2026.

Someone else is doing the talking. The person on video is real. The voice is theirs. But a second person off-camera is feeding them answers in real time, often with an AI assistant generating the responses. The candidate essentially reads what their coach types. It works until a surprise question breaks the rhythm.

The face on screen isn’t real. Open-source deepfake tools now generate convincing video overlays in real time. The person applying may be entirely different from the person who shows up to work. Reports have surfaced of candidates who passed every interview and then arrived on day one as someone completely unrecognizable.

The identity doesn’t match the resume. A single polished resume gets shared across a group of applicants, each of whom interviews under the identity of the strongest candidate. The best interviewer in the group lands the job. The person who actually shows up to work has none of the qualifications claimed.

Remote and hybrid roles are the most exposed. Industries hiring at high volume are being targeted deliberately.

What to actually do about it

You don’t need to turn every interview into airport security. A few practical additions catch most of what’s happening without changing the experience for legitimate candidates.

Add live, unscripted interaction. A quick “before we start, can you hold up today’s newspaper or your phone screen with the date on it” catches most video overlays instantly. A real candidate does it in a second. A deepfake doesn’t.

Switch cameras or ask for a brief walk. “Can you grab something from your desk” or “show me the view out your window” breaks the frame that most deepfake overlays depend on. Legitimate candidates don’t mind. Fraudulent ones can’t comply.

Vary your questions. If every first-round interview at your company asks the same five questions, candidates can rehearse against them with AI. Mix in questions that force genuine thinking. The pauses tell you more than the answers do.

Verify identity at offer stage, not day one. Many companies catch deepfake candidates for the first time on their first day of work. That’s two weeks too late. A short verified video call or in-person meeting before the offer is signed costs almost nothing and eliminates most of the risk.

The bigger picture

The hiring process was built on the assumption that the person interviewing is the person you’re hiring. That assumption no longer holds automatically, and the companies most exposed are the ones who haven’t yet accepted that it’s changed.

You don’t need to be paranoid. You do need to be deliberate. Every protection above takes under a minute to implement, and together they close most of the gap.

The candidate on your screen may not be real. The hiring decision has to be.

Confident the candidate you’re hiring is the person who’ll show up?

At Exact Staff, we meet every candidate ourselves before they ever reach your interview. Face-to-face. Identity verified. References checked. In a market where fraud is rising, that layer of trust is becoming the most valuable part of what we do.

Contact your Exact Staff Representative to schedule a consultation.