AI interview cheating statistics 2026: what the data really shows
The 2026 numbers on candidates using AI in interviews, from the UK, the US and one large platform study, with who measured what and how far each figure can be trusted.
The Candit Team, 2026-10-20
AI interview cheating statistics for 2026 point the same way: it is common, and hiring teams know it. In the UK, 66% of graduate employers report undeclared AI use in interviews. In the US, 60% of hiring managers say AI has made virtual interviews harder to trust. One platform study flagged 38.5% of candidates in nearly 20,000 automated interviews.
Those three figures measure different things, though, and they're often quoted as if they didn't. Below is each number with who measured it, how, and what it can and can't tell you.
Key AI interview cheating statistics for 2026
| Statistic | Who measured it | Sample | Source |
|---|---|---|---|
| 66% of graduate employers report undeclared AI use in interviews | Institute of Student Employers (ISE) | UK graduate employers, 2026 | Resultsense, citing The Guardian, Oct 2026 |
| 92% of graduate employers suspect or have confirmed cheating in recruitment, up from 85% a year earlier | ISE | UK graduate employers, 2026 | Resultsense, Oct 2026 |
| 60% of hiring managers say AI has made virtual interviews harder to trust, up from 51% | Checkr | 3,000 US hiring managers, Sep 2026 | Checkr via PR Newswire, Oct 2026 |
| Only 23% are extremely confident their process would catch a fraudulent candidate | Checkr | Same survey | PR Newswire, Oct 2026 |
| 38.5% of candidates flagged for likely AI help | Fabric (platform data) | 19,368 automated AI interviews, Jul 2025 to Jan 2026 | Forbes, Aug 2026 |
| 6% of candidates admit to interview fraud, such as posing as someone else | Gartner | 3,000 job candidates, 2025 | HR Dive, Aug 2025 |
How the numbers moved since 2025
Where a survey asked the same question a year apart, every figure went up.
| Statistic | 2025 | 2026 | Source |
|---|---|---|---|
| UK graduate employers who suspect or have confirmed cheating | 85% | 92% | ISE |
| UK graduate employers who give candidates guidance on acceptable AI use | 55% | 82% | ISE |
| US hiring managers who say AI has made virtual interviews harder to trust | 51% | 60% | Checkr |
| US hiring managers who have interviewed someone later found to be using a fake identity | 31% | 40% | Checkr |
| US hiring managers who say candidates fake identities with AI faster than companies can catch it | 62% | 68% | Checkr |
How many candidates use AI in interviews?
Nobody knows the true rate, and any single figure claiming to is overreaching. The closest estimate is the platform study above: 38.5% of candidates crossed its flagging line, an estimated probability of AI help above 40%. That's a flag rate, not a confirmed cheating rate, and the interviews were automated ones run by an AI interviewer rather than live conversations with a person.
The employer surveys measure something else. ISE's 66% is the share of employers who have seen undeclared AI use, not the share of candidates using it. Gartner's 6% is self-reported, and people under-report things they know are against the rules. The honest summary: somewhere between "a few" and "a third" of candidates, depending on the role, the format and who's counting.
Which interviews are most exposed to AI cheating?
Technical interviews, by a wide margin. In the same platform study, technical roles were flagged at 48%, against 12% for sales roles. Junior candidates, with up to five years' experience, were flagged at nearly double the rate of senior ones (Forbes, 2026).
| All candidates | 38.5% |
|---|---|
| Technical roles | 48% |
| Sales roles | 12% |
Fabric, 19,368 automated AI interviews, July 2025 to January 2026. Flag rates, not confirmed cheating.
The study doesn't explain the gap, but it's easy to see why coding and system-design questions suit a copilot: they have answers that can be generated and read out. We look at the reasons in more depth in why technical interviews have become AI cheating's biggest target.
Do hiring teams still trust video interviews?
Less every year. Checkr's 2026 survey found that 60% of US hiring managers say AI has made virtual interviews harder to trust, up from 51% the year before. 40% have interviewed someone later found to be using a fake identity, up from 31%.
The more telling number is confidence. Only 23% are extremely confident their process would catch a fraudulent candidate, and 28% say they knowingly advanced a candidate they suspected of dishonesty because the role needed filling (PR Newswire, 2026). The suspicion is there; the evidence to act on it usually isn't.
Are candidates who use AI getting hired?
Some are getting through. In the platform study, 61% of flagged candidates still scored at or above the pass mark of 7 out of 10. If the flags are right, a score alone would have sent most of them to the next round.
The UK picture points the same way. ISE reports that 82% of graduate employers now give candidates guidance on acceptable AI use, up from 55% a year earlier, which suggests employers are setting rules because they can't rely on spotting it.
How far can you trust these numbers?
Treat every figure here as a direction, not a measurement. Three cautions:
- Vendor data is a vendor's view. The 38.5% figure comes from a detection platform's own flags, with no independent check of which flagged candidates actually used AI.
- Surveys count suspicion. "Suspect or have confirmed" mixes the two, and an employer that saw one case counts the same as one that saw a hundred.
- Employer share is not candidate share. 66% of employers is not 66% of candidates.
Older figures without a traceable source, such as the widely repeated "1 in 3 interviews", are left out of this page on purpose.
What the numbers mean for your interviews
The surveys agree on one thing: hiring teams suspect AI help far more often than they can show it. Lockdown browsers and screen monitoring add friction for every honest candidate and still miss a copilot on a second device. The alternative is to watch the conversation itself.
That's how we vet candidates for clients. We run the live interview ourselves, and our anti-cheat reads how answers are delivered rather than what's on the candidate's screen, noting the exact moment worth probing. That includes the bullet-point format many copilots now use so candidates can put answers in their own words: the pause while the points appear, points delivered in list order, and written phrasing in a spoken answer. There's nothing for the candidate to install.
A flag is a prompt for a follow-up question, never a verdict, and people make every decision. Our guide to telling if a candidate is using AI in an interview covers what to ask next.
Frequently asked questions
Is AI interview cheating getting worse?
Every figure that can be compared year on year went up. 92% of UK graduate employers now suspect or have confirmed cheating, against 85% a year earlier, and 60% of US hiring managers say AI has made virtual interviews harder to trust, against 51%. Fabric also reports that its flag rate rose sharply between July and September 2025 and stayed high through January 2026.
Is using AI in a job interview always cheating?
No. It depends on what the employer allows. 82% of UK graduate employers now tell candidates which AI use is acceptable, and 64% accept AI somewhere in their hiring process, up from 46%. What employers report as misconduct is undeclared use: AI writing the answers the interview is meant to test.
Which AI interview cheating statistic should I quote?
Quote each one for what it measures. ISE's 66% is the share of UK graduate employers who have seen undeclared AI use in interviews. Checkr's 60% is how many US hiring managers find virtual interviews harder to trust. Fabric's 38.5% is the share of candidates one vendor's system flagged in automated interviews. None of them is the share of candidates who cheat.
How can you tell if a candidate is using AI in a live interview?
Watch the delivery rather than the screen: a pause before an unusually polished answer, points delivered in list order, written phrasing in a spoken answer. Then ask a follow-up that only their own experience can answer. We do this in every live interview we run for clients, checked with our anti-cheat, and a flag is always a reason to ask, never a verdict.
Turn suspicion into evidence
The statistics will keep climbing, and no process will make an interview cheating-proof. What you can change is whether you find out during the call or after the offer.
We vet every candidate down to the interview before you meet them. Book a call, or if you already have candidates, we'll check them.