The monthly gut-check for people who hire humans for a living. Which, it turns out, is getting harder and harder to verify.

Let’s start off with an easy question: how confident are you that everyone in your pipeline actually exists?

We’re not talking about candidates who aren’t qualified. We’re not talking about ones that aren’t “culture fits,” whatever the hell that means this quarter. We’re talking about, you know, whether or not the candidates in your funnel exist. 

You know – like, those are real resumes from real people, representing real work histories, real experiences and expertise, and real intelligence – or, at the very least least, a real pulse rendered without having to use a GPU. 

Turns out, that easy question is really anything but – although most recruiters already know by now that the data out there suggests that fake candidates are becoming an increasingly real issue in recruiting and hiring. Gartner predicts that within the next 2 years, one in four candidate profiles worldwide will be fake.

And that’s not even including erstwhile fake profile factory LinkedIn and its self-reported 1.3 billion monthly users. Nor is it talking about embellished resumes, invented skills or disingenuous or misleading interview responses. That’s what “fake” candidates used to look like in talent acquisition. Not today.

Today, fake candidates are just that – “applicants” who, in fact, are completely synthetic, totally made up or (in extreme cases) belonging to somebody who isn’t the person who goes through your hiring process but ends up on your payroll. 

Last month, we discussed whether or not AI was generating real ROI or measurable business impact. This month, we’ll be looking at whether or not your next hire is an actual person, artificial intelligence, or a complete work of fiction.

That should be a pretty simple and straightforward question to answer. But anyone in talent acquisition today knows it’s anything but. At least if you believe the hype and the headlines.

Which raises the question we wanted to explore this month here at Recruiting Unfiltered:

Are fake candidates a real problem for real recruiters?

Faking It to the Limit

In a recent survey of over 3,000 candidates, only 6% reported that they had attempted any type of interview fraud, defined as either posing as someone else, or having someone else pose as them. 

Given the heavy investment existing and emerging technology providers have made in integrating fraud detection into their AI interviewing solutions, and the resources employers are dedicating to combating this purported problem. 

Although the true number might be a point or two higher, the good news for employers is that, for the most part, you in fact do know if the candidates in your pipeline are real people (to a strong degree of statistical certainty). The bad news? What’s now a relatively minor problem is becoming a lot more common.

At least, if you ask employers.

According to an April survey of enterprise technology and information security leaders, fully 41% – or over 2 out of every 5 CIOs or CISOs surveyed – confirmed that their organization has already hired, onboarded or contracted with a fraudulent candidate. 

These fake employees were not only interviewed, but fully onboarded. The thought of a synthetic “worker” getting shipped their own company laptop, credit card, credentials and access to stuff like Slack and Sharepoint is basically nightmare fuel for most businesses – and most businesses, empirically, seem stuck in a bad dream.

So, who – or rather – what – are these fake candidates, exactly? That definition seems somewhat broader, which makes sense, considering there are basically no remaining barriers to entry for anyone looking to get in on the applicant fraud game. 

In a recent study, Palo Alto Networks found that someone with basic computer skills and no previous experience with digital photography or image manipulation were capable of building a fake candidate capable of passing a video interview in just over an hour. That’s less time than most of the interviews these fictitious applicants were programmed to pass.

The implications for the bigger business and bottom line are enormous; the FTC reported that job related fraud losses climbed from a relatively paltry $90 million in 2020 to over half a billion dollars in 2024, the most recent year for which data is available. 

That’s an almost 500% increase in just four years – the kind of growth curve most of the vendors selling services and solutions designed to turn the tide on this trend would kill for.

And the money tells the same story. The FTC tracked job-related fraud losses climbing from $90 million in 2020 to over $501 million in 2024. That’s a 457% increase in four years, which is a growth curve most of the vendors on your shortlist would kill for.

The Click Bait

Here’s where it stops being a quirky recruiting problem and starts being a security briefing.

The headline villain is the North Korean remote IT worker scheme, which the Department of Justice has tied to hundreds of US companies.Staffing Industry Analysts reported this week that an Amazon executive disclosed the company blocked more than 1,800 suspected North Korean applicants, with attempts rising 27% quarter over quarter.

 The FBI has issued advisories specifically warning employers and their staffing suppliers, because contract IT work turns out to be the preferred side door.

But state actors are honestly the exotic tail of the distribution. The bulk of it is more mundane and, in its own way, more depressing. You’ve got proxy interview rings, where one skilled person interviews on behalf of a rotating cast of less skilled clients. 

You’ve got the overemployed crowd collecting multiple remote salaries and outsourcing the actual work. And you’ve got the softest category of all: real people, genuinely themselves, who lean on chatbots so heavily in interviews that their answers reflect approximately nothing they know. 

A recent study found91% of US hiring managers have encountered or suspected AI-generated answers in video interviews.

Ninety-one percent. At that point it’s not an anomaly, nor a trend; it’s the new recruiting reality. 

Ecosystem Cooler: What’s Haunting Hiring

Can we be honest about how we got here, though? The industry spent a decade optimizing for frictionless application processes. One-click apply, Easy Apply, quick apply, apply-with-your-face. We treated every step between “saw the job” and “submitted” like a leak in the funnel and caulked accordingly.

Then AI came along and did what AI does, which is take an existing behavior and scale it past the point of meaning. Joveo’s ownRecruiting Benchmarks Report found applicant volume up as much as 9x over three years for many roles. Somewhere inside that avalanche, an increasing share of applications now comes from bots, proxies, or personas assembled in an afternoon.

So we built a system where applying costs nothing, then acted shocked when people who don’t exist started applying. The candidates automated their side. The employers automated theirs. 

And now, both sides are spending real money to figure out if anyone on the other end of the transaction is human. It’s the recruiting equivalent of two chatbots earnestly negotiating a car lease, except the car is a six-figure salary with health benefits and VPN access.

Gartner’s numbers show the trust rot goes both ways, by the way.Only half of candidates believe the jobs they’re applying to are real, and only 26% trust AI to evaluate them fairly. 

Ghost jobs on one side, ghost candidates on the other. Turns out, the funnel is haunted from both directions.

Who You Gonna Call? 

The temptation right now is to buy a deepfake detector, declare victory, and move on. Fight that urge; detection tools are already in an arms race they’re not obviously winning, and single point solutions are how this industry ends up with 47-tool tech stacks that still can’t answer basic questions.

The more durable moves are structural, and most of them are boring:

  • Verify identity early, not at background check. Traditional background screening happens after the hiring decision is basically made, and it relies on identity data the candidate provided. If the identity itself is fabricated, you’re rigorously verifying a fiction.
  • Bring back some friction on purpose. Google, Cisco, and McKinsey have all reintroduced in-person interviews for exactly this reason. And here’s the twist nobody saw coming: Gartner found62% of candidates say they’re more likely to apply when the process includes an in-person interview. Real candidates want proof the job is real, too.
  • Use structured screening that’s hard to outsource. Live, adaptive, structured interviews with consistent scoring are meaningfully harder to game than the async video responses candidates can rehearse with a chatbot, or hand off entirely. This is a big part of why we builtJoveo’s AI Interviewer the way we did, and why thebuying guide question that matters isn’t “does it have AI” but “can you audit how it scores.”

None of this makes fraud impossible. It makes your pipeline a worse target than the next company’s, which, in security as in bear attacks, is usually enough.

The Recruiting Rundown

Here are some of the best resources for taking a deep dive into candidate fraud we’ve read this month, in no particular order:

Kyle & Co. benchmark candidate fraud researchThe full report argues candidate fraud is now an enterprise trust problem spanning TA, Security, HR, Legal, and IT, and the live executive session on July 29 asks the only question that matters, which is who actually owns the response. Spoiler alert, right now it’s usually nobody.

Gartner on candidate trust and fraud. The source document for the “1 in 4 fake profiles” stat everyone’s quoting, plus the less discussed finding that offer acceptance rates dropped from 74% to 51% in two years. Candidates trust us about as much as we trust them, which is to say, not much. And it’s only getting worse.

Staffing Industry Analysts on compliance angle. If you run a contingent workforce program, this one’s required reading. Speed-to-fill pressure plus remote onboarding equals maximum exposure, and the FBI is now naming staffing suppliers directly in its advisories.

The Interview Guys’ explain candidate deepfakes A genuinely useful breakdown of how the fraud taxonomy actually works, from state-sponsored operations to proxy rings, and why the North Korea angle gives some teams a false sense of what to watch for.

Joveo’s guide to AI screening tools. We compared ten platforms and, more importantly, explained why the categories aren’t interchangeable. If your evaluation process currently consists of “which demo was shiniest,” start here.

Trust Is the New Time-to-Fill

For twenty years, this profession has measured itself on speed and volume. Faster funnels, bigger pipelines, lower cost per application. Every metric we celebrated assumed the thing flowing through the funnel was real.

That assumption just kinda died, and most orgs haven’t updated the dashboard.

The next competitive advantage in talent acquisition probably isn’t another point of automation. It’s verified signal. Knowing that your applicants are human, your data is clean, and your process can prove both. The companies that figure this out will hire slower than the bots and better than everyone else, which is a trade most CFOs will happily take once the first fraudulent hire walks off with the customer database.

The machines are applying. The machines are screening. Someone in this transaction should probably still be a person, and more often than not, that someone is you.

Happy hunting.

Matt Charney for Team Joveo

Want to know what a real applicant pool looks like before you spend against a fake one? TheJoveo Recruiting Benchmarks Report 2026 breaks down volume, cost, and competition by occupation and state. Because talent intelligence beats artificial intelligence, and always will.

Recruiting Unfiltered is published monthly by Joveo. Forward this to someone in talent acquisition who needs to hear it.