I've recently been digging into candidate fraud in recruiting, and the numbers are shocking. In a HYPR survey of 500 U.S. HR leaders published in September 2026, 98% said they had come across fake candidates.
This is no longer an edge case. It is part of the hiring funnel — and for technical roles, where a new hire gets access to code, data and production systems on day one, it is a security problem as much as a recruiting one.
What the data shows
Three findings from the HYPR research stayed with me:
- 98% of HR leaders have encountered candidate fraud.
- In 42% of organisations, hiring fraud is detected only after the employee's first day — when the person already has credentials and access to your systems.
- About a quarter (24%) spend one to three months fully resolving a single fake hire.
And most fraudulent hires are not caught by tools at all. According to the same research, the majority are uncovered through human observation after they passed screening: a manager notices that the person on camera doesn't match the person who interviewed, or the work doesn't match the experience.
What candidate fraud looks like in 2026
Hiring fraud has moved well beyond an inflated resume. The patterns I see discussed most often:
- Fake identities. A real-looking profile, a stolen or synthetic identity, and a person who is not who they claim to be.
- Proxy interviews. One person interviews, a different person shows up to work.
- Deepfake interviews. Real-time face and voice manipulation on video calls, getting cheaper and more convincing every month.
- AI-generated profiles and resumes. Polished LinkedIn profiles, perfectly tailored resumes and invented work histories, produced at scale.
- Interview assistance. Hidden AI tools feeding answers during live technical interviews.
- Account and payroll fraud. Hires who exist mainly to get equipment, access or a paycheck routed somewhere else.
Remote and hybrid hiring made all of this easier. Generative AI made it scalable.
Why AI and engineering teams are a prime target
If you are hiring ML engineers, AI researchers or infrastructure engineers, you are an attractive target. These roles are usually remote-friendly, well paid and come with privileged access — to model weights, training data, customer data, cloud accounts and source code.
A fake hire in a sales role is expensive. A fake hire with access to your production environment or your models can be a breach.
The gap: we verify skills, not identity
Here is what stays with me most: identity checks should happen before the offer, not after.
Most hiring processes go deep on verifying skills and experience — technical screens, take-home assignments, system design interviews, reference checks. But identity rarely gets the same attention. It is often left to a background check after the offer is signed, or to onboarding paperwork on day one.
That made sense when the main risk was an exaggerated resume. With deepfakes and AI-generated profiles getting better, the gap between "we verified this person can do the job" and "we verified this person is real" is getting more expensive.
Red flags recruiters and hiring managers should watch for
None of these prove fraud on their own, but together they are worth a closer look:
- A LinkedIn profile that is new, thin or has few real connections for the seniority claimed.
- Work history that can't be confirmed — no colleagues, no public work, companies that are hard to verify.
- Camera always off, "technical problems" with video, or lip movements and audio slightly out of sync.
- Noticeable differences in voice, accent or knowledge between interview rounds.
- Long pauses before answers, eyes moving to read off-screen, answers that sound scripted.
- Reluctance to do a short unscheduled video call, or to show ID live.
- Requests to ship equipment to a different address than the one on file, or unusual payroll requests.
- References who can only be reached by messenger or personal email.
For technical roles, one of the strongest signals is consistency: does the person who explains their past project in depth sound like the same person who solved the coding problem?
How to build identity verification into your hiring process
A few practical steps I recommend to AI startups:
- Verify identity before the final round, not after the offer. Make a document and liveness check a standard step for every finalist, so it's routine rather than an accusation.
- Keep at least one live, camera-on conversation with the hiring manager — ideally unscheduled or with a follow-up deep dive into the candidate's own past work.
- Check the trail of real work. For engineers that means code, papers, talks and colleagues who can confirm they worked together. This is exactly where research-driven sourcing helps: candidates found through their actual work are much harder to fake.
- Re-verify at onboarding. The person who shows up on day one should be the person who passed the check.
- Limit access in the first weeks. Grant production and data access gradually, and coordinate with your security team.
- Train interviewers. Give hiring managers a short list of red flags and a clear path to escalate concerns.
Anti-fraud tools worth a look
A few identity verification solutions built for hiring:
- Persona — Candidate Verification, launched in March 2026, confirms that applicants are real people at key stages of the hiring process.
- Socure — Workforce Verification runs quiet background risk checks and only asks for a document and biometric check when something looks off.
- ID.me and First Advantage — recently teamed up on an FCRA-compliant identity fraud check for hiring.
The right choice depends on your hiring volume, your ATS and where in the funnel you want to add friction. The important part is deciding where identity gets verified — and making sure it happens before someone gets the keys.
FAQ: candidate fraud
What is candidate fraud?
Candidate fraud is when a job applicant misrepresents who they are or what they can do — from fake identities and proxy interviews to deepfake video calls and AI-generated resumes.
How common are fake candidates?
Very. In HYPR's 2026 survey of 500 U.S. HR leaders, 98% said they had encountered candidate fraud.
When should identity verification happen in hiring?
Before the offer. Verifying identity for finalists, and again at onboarding, closes the gap that lets fraudulent hires get access to systems.
How can you spot a deepfake interview?
Watch for audio and lip movements out of sync, unnatural facial movement when the person turns or touches their face, camera problems, and inconsistencies between interview rounds. A short live, unscheduled video call is one of the simplest checks.
The takeaway
Candidate fraud is now a standard hiring risk, and AI is making it harder to spot. Keep verifying skills deeply — but give identity the same attention, and move it before the offer.