How to Detect Candidate Fraud in Remote Hiring: 12 Best Practices
Fake identities, stand-in interviewees, deepfake video and AI-written answers are now part of remote hiring. Twelve practical checks, from the CV to the first day, that catch most candidate fraud without treating every honest applicant as a suspect.
Ece Badak • Oct 17, 2026 • 5 min read
Remote hiring made it possible to hire anyone, anywhere. It also made it possible for someone to apply as anyone, from anywhere.
The numbers are no longer small. In a 2025 Gartner survey of 3,000 job candidates, 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 2025).
Some of this fraud is organized. In June 2025, the US Department of Justice announced coordinated actions against schemes in which North Korean IT workers used stolen or fake identities to get remote jobs at more than 100 US companies, including searches of 29 suspected "laptop farms" across 16 states (US Department of Justice). As early as 2022, the FBI warned that deepfake video and stolen personal information were being used to apply for remote jobs (FBI IC3).
This guide covers the common types of candidate fraud and twelve practical checks against them.
The four common types of candidate fraud
- Embellished or AI-written CVs. The most common type. The CV copies the job post's keywords and describes experience the person does not have.
- AI help during the interview. The candidate reads answers generated in real time by an AI tool on the same screen or a second device.
- Proxy interviews. Someone else does the interview, sometimes with the real candidate lip-syncing, and a different person starts the job.
- Fake or stolen identities. The person is not who they say they are at all. This is the type with the highest security risk, because the goal is often access to systems and data, not the salary.
Each type needs a different check. No single tool catches all four.
Before the interview
1. Read the CV for duties, not keywords. A fake CV usually matches your job post well, because it was written from it. Look at what the person did in each role: team size, tools used for what, results. A CV that repeats your requirements without saying what the person did is a signal to probe, not proof of fraud.
2. Check the timeline against the public record. Do the employers exist? Do the dates and titles match the candidate's professional profiles? Is a profile with ten years of experience only a few months old, with few real connections? One mismatch is normal. Several together deserve a question.
3. Watch the contact details. Voice-over-IP phone numbers, email addresses created recently, and requests to send equipment to an address different from the stated home address are common in organized schemes.
During the interview
4. Ask follow-up questions about the candidate's own work. This is the single most effective check. A real story gets more detailed when you ask "What exactly did you do?", "What went wrong?" or "What would you do differently?" A memorized or AI-generated answer gets vaguer, or the candidate pauses before every reply.
5. Watch for patterns, not single moments. Common signs of real-time AI help include a regular delay before each answer, eyes that move as if reading, polished but generic answers, and answers that fall apart under a follow-up. Any one of these can be nerves. Several together, repeatedly, are worth a closer look.
6. Keep the camera on and the screen visible. For remote interviews, ask candidates to keep the camera on and, where the format allows, to share their screen. Tell them in advance, and explain why, so honest candidates do not feel accused.
7. Ask something unexpected and local. A question about the city they live in, their commute or a detail from a past employer's office is easy for the real person and hard for someone working from a script.
8. Compare across interviews. Is it the same voice, the same face and the same level of knowledge in every round? Proxy interviews often show up as a strong first round and a much weaker second, or the reverse.
After the offer
9. Verify identity with a document check before the start date. Match the identity document to the person you interviewed, by live video if the role is fully remote. In regulated industries this is usually required anyway.
10. Ship equipment only to a verified address. Requests to send a laptop to a third party's address are one of the clearest signs of an organized scheme.
11. Check that the person on the first day is the person you interviewed. A short live video call with the hiring manager in the first week catches proxy hires early.
12. Limit access in the first weeks. Give new remote hires the access their first tasks need, not everything at once. If something is wrong, the damage stays small.
What not to do
- Do not rely on "AI detectors" for CVs. They produce false positives, and many honest candidates now use AI to polish their writing.
- Do not treat every non-native accent or a slow answer as suspicious. Fraud checks must apply the same rules to everyone, or they become discrimination.
- Do not use facial analysis to judge honesty. It is unreliable, and in the EU, AI systems that infer emotions in the workplace, which includes recruitment, are banned under the AI Act.
- Do not hide your checks. Tell candidates what you check and why. Gartner's research also recommends explaining your rules on acceptable AI use. Honest candidates accept fair rules; fraudsters are discouraged by them.
How AI interviews change the picture
AI interviews can make fraud harder or easier, depending on how they are built. A one-way video tool that only records answers gives a fraudster unlimited time and no follow-up questions. A live interview with integrity rules is much harder to fake.
Firstview's AI Interviewer asks follow-up questions in real time when an answer is thin, and Firstview's interview proctoring applies live rules while the interview runs: the whole screen is shared, the interview stays in full screen, the camera stays on and virtual cameras are blocked. When a rule breaks, the interview pauses and the candidate gets 4 minutes to fix it. Every candidate gets the same rules.
To be clear about the limits: Firstview does not verify identity documents and cannot see a second device such as a phone. It reduces cheating; it does not prove that none happened. That is why the checks before and after the interview in this guide still matter.
