How to spot fake candidates in remote technical interviews
Proxies, deepfakes and stolen identities now reach live interviews. A stage-by-stage checklist for recruiters screening remote engineers, with 2026 data.

On this page
- How common are fake candidates in 2026?
- What do fake candidates look like in a technical process?
- Where do fake candidates slip through?
- A stage-by-stage checklist for recruiters
- How to make the technical round hard to fake
- How to check identity without treating every candidate as a suspect
- Where an AI screen helps, and where it doesn’t
- Frequently asked questions
- Sources
Fake candidates now get through remote technical interviews often enough that most hiring teams have met one. The pattern is consistent: a different person or a synthetic face takes the interview, a stolen or invented identity carries the paperwork, and the fraud surfaces only after a laptop ships. The defense that works is a process: verify identity before the first live round, keep the same person accountable at every stage, and make the technical interview hard to fake in real time.
How common are fake candidates in 2026?
Common enough that it is now a routine hiring risk, not a security team’s edge case. Checkr surveyed 3,000 US managers who had been involved in hiring in the previous 12 months, fielding the survey September 11-16, 2026. 40% said they had personally interviewed a candidate who was later revealed to be using a fake identity. 67% had confirmed or strongly suspected an interview proxy (Checkr, The Hiring Hoax 2026). An interview proxy is a person who takes an interview in place of the real candidate, usually because they are the stronger engineer.
Confidence has not kept up. In the same survey, only 23% of managers were extremely confident their process would catch a fraudulent candidate. 68% believed job seekers are now better at faking identities with AI than hiring teams are at catching them.
The candidate side tells the same story. In a Gartner survey of 3,000 job candidates in the second quarter of 2025, 6% admitted to interview fraud, either posing as someone else or having someone else pose as them. Gartner predicts that by 2028 one in four candidate profiles worldwide will be fake (Gartner, July 31, 2025). That is a forecast, not a measurement, but it shows where the analysts think the curve is going.
The organized end of the problem is targeting technical roles. On October 13, 2026, the Justice Department announced charges against 14 people in a North Korean remote IT worker scheme that placed workers at more than 70 US companies and brought in about $3 million. In the same action, law enforcement searched more than 40 residences in Texas that were known or suspected laptop farms, and seized about 187 laptops and KVM devices (US Department of Justice, October 13, 2026). A laptop farm is a home where a facilitator keeps company-issued laptops switched on so that workers abroad can log in remotely as if they were in the United States.
| Source | Finding | Date |
|---|---|---|
| Checkr survey, 3,000 US hiring managers | 40% interviewed a candidate later found to be using a fake identity (31% in 2025) | Oct 2026 |
| Checkr survey | 67% confirmed or strongly suspected an interview proxy | Oct 2026 |
| Checkr survey | 23% extremely confident their process would catch fraud | Oct 2026 |
| Gartner survey, 3,000 candidates | 6% admitted to interview fraud | Q2 2025 |
| US Department of Justice | 14 charged; more than 70 US companies affected; about 187 devices seized | Oct 2026 |
If you hire for clients, this lands on you first. Our earlier look at what hiring leaders said about AI agents applying to jobs found a similar shift on the employer side: more applications, more misrepresentation, and more trust placed in live, verifiable signals.
What do fake candidates look like in a technical process?
Four patterns account for most of what recruiters describe. They often appear together.
- The proxy interviewer. A skilled engineer takes the screen and the technical rounds; a different person shows up for the job. A video call with a decent camera hides most of the differences between two people.
- The real-time face swap. Deepfake software maps a different face, and sometimes a different voice, onto the person on the call. It is most convincing when the candidate stays still, faces the camera and the call quality is poor.
- The borrowed or synthetic identity. A stolen identity or a stitched-together one carries the résumé, the background check and the payroll paperwork. This is the model in the North Korean IT worker cases, where US-based facilitators receive company laptops and keep them online for remote operators.
- Live assistance. The candidate is real but reads answers generated by an AI tool or fed by someone off camera. Whether this counts as fraud depends on your rules, which is why you should state them in advance (more on that below). It also explains why a polished résumé proves less every year, a shift we covered in why AI-polished résumés became recruiters’ top stress.
The common thread: each pattern exploits a step where the hiring process trusts what it sees without connecting it to the next step.
Where do fake candidates slip through?
At the handoffs. Checkr found that only 29% of companies verify identity as soon as someone applies, while 49% wait until after at least one live interview. And 63% concentrate their fraud prevention before the hire, against 6% after it (Checkr, 2026).
That leaves two gaps. First, a proxy can pass the recruiter screen and the technical round before anyone checks who they are. Second, nobody compares the person at onboarding with the person who interviewed. The Justice Department case shows why the second gap matters: the payoff for the scheme starts when the company ships a laptop to an address the operator controls.
A stage-by-stage checklist for recruiters
The checklist below is our synthesis of the Justice Department case, Gartner’s recommendations and the Checkr data, organized by hiring stage so each step has an owner. Use the same checks for every candidate at the same stage. Consistency is what makes the process fair, and it is what makes anomalies stand out. Identity verification here means a document check plus a live selfie match, run by a verification vendor rather than by the recruiter.
| Stage | What to check | Red flags worth a second look |
|---|---|---|
| Application | Does the identity hold together across résumé, LinkedIn, GitHub, email and phone? Profile age, real connections, commit history that matches the claimed experience | A brand-new LinkedIn profile with senior titles; the same phone number or email pattern across “different” applicants; a GitHub account with no history before this year |
| Before the first live round | Tell candidates up front that interviews are on camera without virtual backgrounds and that identity will be verified. Run document-plus-selfie verification for everyone who reaches this stage | Refusing verification, or asking to “do it later”; documents whose name or photo doesn’t match the profile |
| Recruiter screen | Ask about specifics only the real person knows: the team they were on, what they shipped, why they left. Confirm time zone and location against what they wrote | Answers that stay generic under follow-up; a time zone or background that doesn’t match the stated location; long pauses before every answer |
| Technical interview | Keep it live and interactive: change a requirement midway, ask why they chose an approach, ask them to walk through code they wrote. Record with consent so later rounds can compare | Strong code with no ability to explain it; eyes tracking off-screen text; audio that lags the lips; a different voice or manner than the screen |
| Offer and background check | Verify past employers through phone numbers you find yourself, not the ones the candidate supplies. Match the name, photo and documents to the person you interviewed | References only reachable through numbers the candidate gave; employment you cannot confirm independently |
| Onboarding | Ship equipment only to the verified address. Open day one on video with the same person. Ask IT to watch for remote-access tools and logins from unexpected locations | Last-minute shipping address changes; a request to ship to a “relative”; KVM or remote-desktop software on a new laptop |
Two rules make the checklist work. Keep one record per candidate that follows them from screen to onboarding, so each stage can be compared against the last. And name an owner for the handoff to IT, because the recruiter often leaves the process exactly when the laptop goes out.
How to make the technical round hard to fake
A take-home exercise proves that someone can produce the code; it does not prove who produced it. A live round can, if it is designed to require thinking out loud.
- Change the problem midway. Add a constraint (“now it has to handle a million records”) and ask what changes. Scripts and stand-ins struggle with changes they did not prepare for.
- Ask about their own decisions. “Why a hash map here?” or “What would break first in production?” are easy for the person who wrote the code and hard for someone reading a feed.
- Use spoken follow-ups, not just code. Voice answers under follow-up questions reveal understanding in a way that a pasted solution doesn’t. A bank of real interview questions helps keep rounds consistent between candidates.
- Keep the same evaluator and rubric across rounds. If the person in round three answers differently than the person in round one, a consistent rubric makes that visible.
For example, a candidate writes a clean in-memory rate limiter. Ask them to make it work across three servers. An engineer who wrote it starts reasoning out loud about shared state, clocks and what happens when a node fails. Someone reading a feed tends to restate the original solution or go quiet until a new answer appears.
Recruiters screening backend or platform engineers can adapt the follow-ups to the role; our role hubs show what current postings ask for.
How to check identity without treating every candidate as a suspect
Most of your candidates are real, and they notice how they are treated. Gartner’s advice to recruiters is to set expectations early: explain which uses of AI are acceptable and that fraud is detected and has consequences (Gartner, 2025). Honest candidates generally accept friction that is explained. In Gartner’s survey, 62% of candidates said they were more likely to apply to a role that requires in-person interviews.
A few practices keep the process fair:
- Same check, same stage, every candidate. Never trigger extra checks because of an accent, a name, a camera that won’t focus or a bad connection. Those are not fraud signals.
- Explain what you collect and why, and keep identity documents with your verification vendor rather than in email threads.
- Separate “AI help” from “fraud” in writing. If candidates may use an AI assistant to prepare but not during the live round, say so in the invitation.
- Have employment counsel review the process, especially if you hire across states or countries with different rules on background checks and automated decisions.
Where an AI screen helps, and where it doesn’t
A structured AI screen gives every candidate the same interview, records it, and produces a report you can compare with later rounds. Prepin works this way for technical roles. The candidate gets one link, an AI interviewer runs a voice technical interview with live coding, and the recruiter gets a scored report before anyone books a call. Consistent questions and a recording make a mismatch between the screen and the onsite easier to spot. A screen does not verify identity on its own, though, so pair it with the document check above. You can see how screening works for agency recruiters.
Frequently asked questions
What is an interview proxy?
A proxy is a person who takes an interview on behalf of the actual candidate, usually because they are more skilled. The candidate who later starts the job is a different person. In Checkr’s 2026 survey, 67% of US hiring managers had confirmed or strongly suspected a proxy.
How can you tell if a candidate is using a deepfake on video?
Look for audio that drifts out of sync with the lips, edges of the face that blur when the person moves, and reluctance to turn sideways or move a hand across the face. None of these is proof on its own. Identity verification before the interview, and comparing the same person across rounds, is more reliable than spotting artifacts.
Should we verify identity before the first interview?
Yes, if you can apply it to every candidate who reaches that stage. Checkr found that about half of companies (49%) wait until after at least one live interview, which is exactly where proxies operate. Tell candidates in advance and use a verification vendor rather than collecting documents yourself.
What should we do if we suspect a candidate is fake?
Don’t accuse anyone on the call. Pause the process, ask for the same identity verification every candidate gets at that stage, and compare the recordings from earlier rounds. Bring in your security and legal contacts before an offer or any equipment goes out. If you suspect a North Korean IT worker scheme, the FBI asks companies to report it through its Internet Crime Complaint Center at ic3.gov.
Is using AI during an interview fraud?
It depends on the rules you set. Using AI to prepare is normal; reading AI-generated answers in a live round that is meant to test the candidate’s own thinking breaks the purpose of the round. State the policy in the interview invitation, so the line is clear before anyone crosses it.
Sources
- Checkr: The Hiring Hoax 2026, survey of 3,000 US managers involved in hiring, fielded September 11-16, 2026
- Gartner: Survey shows just 26% of job applicants trust AI will fairly evaluate them (July 31, 2025)
- US Department of Justice: Coordinated actions to combat North Korean remote IT worker schemes affecting more than 70 US companies (October 13, 2026)
Screen engineers before the first call
Send candidates one link. Prepin runs the technical voice interview and live coding, then gives you a scored report.
See how screening works →Practice the interview out loud
Run a realistic AI mock interview for your target role and get feedback on every answer.
Start a mock interview →