Red 3D wireframe warning triangle with a white exclamation mark against a dark dotted background, representing cybersecurity and remote AI engineer hiring fraud risks.
Back to all articles

40% of Hiring Managers Have Interviewed a Fake Candidate: How Do You Vet Remote AI Engineers Now?

40% of hiring managers have interviewed a fake candidate. Learn how to spot proxy interviews, laptop farms, and AI fraud when vetting remote AI engineers.

Profile photo of Daniela Morales

Daniela Morales

TA Manager , Golabs Insights

Artificial Intelligence (AI)CybersecurityStaff Augmentation

Hiring fraud used to mean a padded resume. In 2026, it means a candidate who may not be the person on the video call. Checkr surveyed 3,000 US managers involved in hiring between September 11 and 16, 2026. 40% said they had interviewed a candidate later revealed to be using a fake identity, and 68% agreed that job seekers are now better at faking identities with AI than hiring teams are at catching them.

The fraud does not stop at the offer letter. 38% of the same managers believe a fraudulent employee is currently working at their company, yet 63% of organizations focus their fraud prevention on the pre-hire stage, and only 6% put the most weight on what happens after someone starts.

The most organized version of this fraud is a national security case. On October 13, 2026, the US Justice Department announced charges against 14 people in a North Korean remote IT worker scheme that reached more than 70 US companies and generated approximately $3 million. The FBI searched more than 40 suspected laptop farms in Texas and seized about 187 computers and KVM devices.

Why Are Remote AI Engineering Roles a Target for Fraud?

Remote AI engineering roles combine three things fraudsters want: high pay, interviews held entirely over video, and access to code, data, and cloud credentials from week one. A fake accounting clerk costs you a salary. A fake engineer with production access can cost you your source code, your customer data, and your standing with regulators.

A note on the evidence. Checkr's survey does not break its results out by role, and we found no independent 2026 data on fraud rates for AI or software engineering hires specifically. What we can say is that the Justice Department's cases involve IT roles by definition, and the motive Checkr's respondents named most often collecting pay without doing legitimate work (41%) is easiest to pull off in a remote role.

Gartner put the risk plainly in its July 31, 2025 research on candidate fraud. Jamie Kohn, a senior research director in Gartner's HR practice, said: "Candidate fraud creates cybersecurity risks that can be far more serious than making a bad hire." Gartner also predicted that by 2028, one in four candidate profiles worldwide could be fake. Treat that as a forecast, not a measurement, but the direction matches what hiring managers report today.

What Does Candidate Fraud Look Like in 2026?

It is more than fake identities. Checkr's respondents described several kinds of deception, and most of them grew over the past year. Each one calls for a different check, which is why a single control, such as an ID scan at the application stage, catches only part of the problem.

Proxy interviews. 67% of managers have encountered interview proxy fraud, up from 35% a year earlier. Of those, 34% confirmed a proxy and 33% strongly suspected one. The person who passes the technical interview is not the person who shows up to work.

Deepfake and AI-generated interview content. 38% reported AI-generated or deepfake content in interviews in the past year, and 60% said AI has made virtual interviews harder to trust, up from 51%.

Misrepresented skills. 65% have uncovered a candidate who misrepresented skills or experience, up from 60%. In the AI candidates we screen, this is a frequent problem: a candidate who lists production machine learning work but has only completed tutorials.

Hidden second jobs. 58% of managers believe an employee is holding another full-time job, and 29% have confirmed at least one case. That is less dangerous than a state-sponsored operative, but you are still paying for attention you do not get.

Organized identity schemes. The North Korean operations go furthest. According to the Justice Department, workers used stolen or fabricated identity documents, VPNs, and proxy infrastructure to appear US-based. Facilitators in the US received company laptops at their homes, installed remote-access software, and forwarded the earnings. Earlier FBI warnings, which the release references, describe data extortion and theft of proprietary data from victim companies.

Why Isn't Pre-Hire Screening Enough Anymore?

Because the most damaging fraud shows up after day one. A proxy can pass your interviews, a real person can pass your ID check and then hand the job to someone else, and a laptop can be shipped to a verified address and then rerouted. Checkr found that 44% of organizations lack a formal, written process for verifying identity after hire.

Time matters more than most teams realize. Checkr cites 2026 data from the Association of Certified Fraud Examiners: fraud caught within six months has a median loss of $40,000, while fraud that goes undetected for five years or more has a median loss of over $1.1 million. The ACFE figures cover occupational fraud broadly, not hiring fraud alone, but the lesson about detection speed carries over.

Hiring pressure makes it worse. 28% of Checkr's managers admitted they had knowingly advanced a candidate they suspected of dishonesty because the role needed to be filled quickly. That is exactly the pressure a fraudster counts on.

How Should You Vet a Remote AI Engineer Now?

Verify the person, the skills, and the access separately, and keep verifying after the start date. These are the checks we use when we place AI engineers. None of them requires exotic tools.

Verify identity twice. Check government ID and run a third-party background check before the offer, then confirm on the first day that the person who logs in is the person you verified. Checkr found that 29% of companies verify identity when a candidate applies, 49% after at least one live interview, and 18% only at the final interview or later. Earlier is better, but a second check at onboarding is what catches the swap.

Test skills on work that cannot be outsourced. Take-home assignments are easy to hand to someone else. Ask candidates to walk through a system they say they built: the data pipeline, the evaluation set, the decision they would change now. Then pair with them live on a small problem in your own stack. A proxy can memorize answers, but explaining trade-offs in their own past work is much harder to fake.

Keep one verified touchpoint. Gartner research cited by Checkr found that 72% of recruiting leaders now require in-person rounds to counter candidate fraud. If an in-person meeting is impractical for a remote role, use a proctored session or a meeting with a trusted partner on the ground. Notably, the share of Checkr's respondents relying on real-time video interviews as a fraud check fell from 55% to 42%, showing how much faith hiring teams still have in video alone.

Control where the hardware goes. CrowdStrike's research on the North Korean groups, reported by The Hacker News, describes how operatives redirect company laptops: false police reports claiming a laptop was stolen to trigger a replacement, rotating residential addresses, and last-minute relocation requests during shipping. Ship only to a verified address, and treat any change requested after the offer as a reason to re-verify.

Grant access in stages. Start new engineers with the least access their first tasks require and widen it as they deliver. Watch for logins from unexpected locations, remote-access software on company devices, and work patterns that suggest someone else is doing the job. Gartner recommends risk-based data monitoring and anomaly alerts after hire. Our guide to agent authentication applies the same least-privilege principle to AI agents.

What Should You Ask a Staffing Partner About Vetting?

Ask how they verify, not whether they do. CrowdStrike found that some North Korean operatives got jobs through IT staffing firms, which allowed end-client laptops to be redirected to laptop farms without the staffing firm's knowledge. That should change how every company evaluates a talent partner, including us.

Five questions separate a real vetting process from a promise. Who verified this engineer's identity, when and how? Has anyone at the firm met the engineer, in person or through a trusted local partner? Who ships and holds the company hardware? What happens if identity cannot be reconfirmed after the start date? And how often does the firm re-verify engineers on long engagements?

Where Should You Start?

Start with the engineers you already have. If 38% of managers believe a fraudulent employee is already on payroll, the first step isn't a new screening tool but a quiet review of current remote engineers with broad access: confirm identities, check where company devices are, and review access logs for anything that doesn't fit.

Then write down the process. Name who verifies identity at each stage, who approves hardware shipments, and who owns post-hire checks. A written process is what lets you hire quickly without skipping steps when a role has been open too long. If you are hiring embedded or client-facing engineers, our post on forward deployed engineers covers how to assess them.

Common Questions About Vetting Remote AI Engineers

How common are fake job candidates in 2026?

Common enough that most hiring managers have seen one. Checkr's September 2026 survey of 3,000 US managers found that 40% had interviewed a candidate later revealed to be using a fake identity, up from 31% in 2025, and 67% had encountered proxy interview fraud. Gartner predicts that by 2028, one in four candidate profiles worldwide could be fake.

What is a laptop farm?

A laptop farm is a home where a facilitator receives company-issued laptops for remote workers who are actually somewhere else, then installs remote-access software so they can log in and appear to work from the US. In October 2026, the Justice Department announced searches of more than 40 suspected laptop farms in Texas tied to North Korean IT workers.

How do you detect a proxy in a technical interview?

Ask candidates to explain systems they say they built, including the trade-offs and what they would change, and pair with them live on a small problem in your own stack. Keep at least one verified touchpoint, in person or proctored, and confirm on the first day that the person who logs in is the person you interviewed.

Is hiring through a staffing firm safer than hiring directly?

Not automatically. CrowdStrike found that some North Korean operatives obtained jobs through IT staffing firms, with client laptops redirected without the firm's knowledge. Ask any partner who verified the engineer's identity and when, whether anyone has met them, who controls the hardware, and how often engineers are re-verified on long engagements.

If you are hiring remote AI engineers and want a vetting process you can explain to your security team, talk to the Golabs team. We verify identity, skills, and access at every stage through our AI talent solutions, nearshore staffing, and AI dedicated teams.

Tagged in

Artificial Intelligence (AI)CybersecurityStaff Augmentation

Save this article

Work with Golabs

Turn your next product idea into working software.

Partner with a senior LATAM engineering team focused on delivery, transparency, and long-term outcomes.

Loading related posts...