Four of eight candidates were not who they said they were

A talent acquisition specialist opened a remote job posting for a Polish company and had fifty applications within seconds. She screened eight of them for a first-round call. By the end of the first day, three had turned out not to be Polish at all. The next day, a fourth fake profile showed up.

Thor Hesselberg watched this happen inside his own recruitment firm, Deca People, in Stavanger. “They had built LinkedIn profiles, they had forged companies, and everything looked perfect on paper,” he told me. That screening round is the reason he founded WorkID.ai, a platform that checks a candidate’s identity and work history before the first interview rather than after the last one.

Watch on YouTube

Why a resume that costs nothing to fake broke an industry built on trust

Hiring has barely changed its order of operations in fifty years, Thor argues. First you assess, then you verify, almost always at the very end of the process. That order made sense when faking an application took real effort. It stopped making sense once it didn’t.

Analysts now put the number of people admitting to lying or leaving gaps in their CV as high as 64%. Thor draws a hard line inside that number: stretching a start date by a month or two is not the same as claiming you worked somewhere you never set foot. “You can take a job description, stick it into an AI model, and fabricate a CV in seconds,” he said. The skill to fake convincingly used to be scarce. The cost of faking has dropped to almost nothing, so more of it shows up, not because people got more dishonest, but because the barrier disappeared.

Gartner’s own estimate is that by 2028, one in four candidate profiles worldwide could be fake. Thor’s response to that number, on air, was blunt: “I’m not sure if it will be, but it can be.” He treats it as a real possibility rather than a guaranteed outcome, which is the more useful way to read any forecast about fraud.

What a fake candidate actually costs a hiring team’s calendar

The clearest way to see the cost of hiring fraud is to look at the North Korean operation that sent out 167,000 job applications and got 76 people hired into remote roles. KnowBe4, a security awareness company, disclosed in 2024 that it had hired a fake North Korean IT worker as a software engineer. If a security firm built to spot exactly this kind of risk can miss it, the odds of catching it with a gut feeling and a good interviewer are not great.

Thor’s most unsettling story from the show is the one about a hiring manager who spent forty-five minutes interviewing a candidate for a remote sales role before realizing he was talking to an AI avatar, not a person. “That’s just crazy, isn’t it?” he said, and it is, but it is also where hiring is heading once every step of the process happens on a screen. A digital-first hiring process makes it easy to feed an interviewer answers off a script in real time, and it makes it just as hard for the interviewer to tell.

Every hour spent assessing someone who was never real is an hour lost to hiring efficiency, not because the hiring manager did anything wrong, but because the checks that would have caught the fraud happen too late to save that time. Thor frames every hiring decision as an access decision: whoever gets hired gets your data, your systems, and your trust, and a bad actor who gets in the door with bad intent on the way in rarely improves once inside.

A smartphone showing a security verification lock icon
Photo by Dan Nelson on Unsplash

Verification, endorsement, and reference: three checks hiring collapses into one

The most useful distinction Thor makes is one most hiring tools flatten into a single tickbox. He splits the problem into three separate things.

Verification is the binary fact: did this person work at this company, in this role, for this period. It is the kind of thing an HR department can confirm in minutes, because most companies already have a compliance process for answering exactly that question when a former employee’s new employer calls.

Endorsement is a colleague willing to vouch for someone as a person. It tells you they are hardworking and trustworthy, but not necessarily that they are right for the specific job on the table.

Reference is the one-on-one conversation a hiring manager actually wants before an offer goes out, the confidential call where real judgment happens. Thor is clear that WorkID has not tried to automate this part and does not intend to. “I would never suggest to take out the human judgment in any hiring process,” he said, “because at the end of the day, humans deal with humans.”

Collapsing all three into one check is, in his telling, part of why fakes get through: a system built to confirm one thing ends up being trusted to answer all three.

The trust score, and who it leaves out

WorkID gives a verified candidate a trust score from 0 to 100, something Thor compares to a credit score. That comparison cuts both ways. Credit scores have a long record of locking people out for reasons that have nothing to do with whether they would be a good hire, and a trust score built the same way risks the same blind spots: freelancers with no single employer to vouch for them, founders whose last company no longer exists, anyone whose former manager has a grudge rather than a reference.

Thor’s own numbers from an internal validation round this spring: 86% of companies he spoke to wanted a verified profile early in the pipeline, against 75% of candidates willing to go through the verification effort. He attributes the gap to the fact that candidates have not yet felt the problem the way companies have. Whether that gap closes, and how WorkID treats the people a credit-style score tends to leave out, is the question worth watching as the platform grows past its early clients.

On privacy, Thor says WorkID stores data within the EU, follows GDPR, and builds toward SOC2 compliance, with the records built on a blockchain so they are tamper-proof and owned by the candidate rather than by any employer. The candidate decides who sees the record and can revoke access at any time, at least in the version Thor described on the show.

A laptop open on a video call during a remote hiring conversation
Photo by Chris Montgomery on Unsplash

Proof beats polish, starting this week

Thor’s estimate for a company that moves verification to the first step: roughly 50% of the time currently lost in screening comes back, and the need for a full background check later in the process drops by about 80%, though not to zero, since some roles still legally require it. The bigger shift is qualitative. Recruiters spend less time filtering out people who were never real, and more time on the judgment calls that actually need a person.

For a candidate with no verified profile anywhere, Thor’s practical advice is simple: ask the company for the same thing it is asking of you. “I would actually urge them to say, you ask me for a reference, I want you to give me a reference to somebody working in the company,” he said, framing it as a mutual decision rather than a one-way audit. Some companies in Norway already run a reversed reference check along these lines, where the employer offers the candidate someone to call.

Thor’s own closing advice, when I asked him for his best hire in one sentence, had nothing to do with verification at all: “Hire somebody who is smarter than yourself. That’s the best advice I can give.” It is a reminder that a verified CV only tells you someone is who they say they are. Whether they are the right person is still a human call, every time.

Frequently asked questions

What percentage of job candidates lie on their CV?

Analysts cited on the show put the figure as high as 64% for some form of lying or padding, though most of that is minor, such as adjusting dates to avoid a gap. Thor Hesselberg draws the line at claiming to have worked somewhere you never did, which he treats as a different category entirely.

How many fake job applications came from the North Korean IT worker scheme?

One operation sent 167,000 applications and got 76 people hired into remote roles, a case Thor cites to show the scale fraud can reach once applying is nearly free. KnowBe4 separately disclosed hiring one such fake worker as a software engineer in 2024.

What is the difference between verification, endorsement, and reference in hiring?

Verification confirms the binary facts of someone’s work history. Endorsement is a colleague vouching for them as a person. Reference is the one-on-one conversation a hiring manager has before an offer, the part Thor says should never be automated out of the process.