Using AI to get through a live interview is no longer a fringe trick. In an April 2025 survey of more than 3,600 professionals, one in five admitted to secretly using AI during a job interview, and 55% called it "the new norm". A 2026 Resume Genius report found 22% of candidates have used AI in real time during a live interview, and Greenhouse reported that 36% of job seekers have used AI to alter their appearance, voice, or background on a video call.

Interviewers feel it. A Checkr survey of 3,000 managers found 59% suspect candidates of using AI to misrepresent themselves — but only 19% are confident their process would actually catch a fraudulent applicant. That gap, between suspicion and proof, is the whole problem.

What candidates are actually using

The market has moved well past "alt-tab to ChatGPT."

The uncomfortable part: the polished tools advertise that they stay invisible during screen share. So "ask them to share their screen" is no longer the safety net it used to be.

The tells

No single sign is proof. But these patterns, especially in combination, are worth noting. They come from interviewers and researchers, not guesswork.

  • Eyes drifting off-camera. A candidate reading an answer tracks a fixed line off to the side. Harvard Business Review flags "their eyes keep drifting off camera"; Karat notes frequent glancing off-screen or window switching.
  • A consistent pause, then a suddenly complete answer. A short, repeatable delay — often three to five seconds, long enough for an AI to process — followed by an unusually polished reply.
  • Answers that are too tidy. Templated scaffolding ("There are three main points…"), textbook phrasing, and generic wording that lacks any personalization.
  • They cannot go off-script. SHRM Labs puts it bluntly: "A genuine high performer can explain their thinking. A skillfisher can't." When pushed past the prepared answer, the depth disappears.
  • A reading cadence. Flat, even tone with no natural emphasis or self-correction.
  • No specifics. AI has no memories. Ask for the name of the person in a conflict story or the exact metric on a project and the detail evaporates.

How to test it, live

You do not need a tool to pressure-test an answer. You need follow-ups that AI struggles to keep up with.

Why this is getting harder — and what helps

Most interviewers think they can tell: 88% of hiring managers say they can spot AI use. But confidence is not the same as evidence — recall that only 19% trust their process to actually catch it. As overlays get better at hiding from screen share, the human eye alone is a shrinking advantage, which is why Gartner predicts one in four candidate profiles will be fake by 2028.

The reliable answer is not more suspicion. It is better signals: tying the small tells above — focus changes, paste velocity, answer structure, reading gaze — to the exact moment they happen, so you are acting on a timeline rather than a gut feeling. That is the read Trueyy is built to give you, consent-first and in real time.

The best candidates do not need AI to get through your screen. The goal is simply to make sure the person you hire is the person you talked to.

Sources