A recruiter on Reddit described a moment that's quietly becoming familiar across talent acquisition. A few minutes into a video interview, something felt off, the expressions, the timing, the way the face moved. Then it clicked: the person on screen wasn't real. They were running an AI avatar.
AI interview fraud has moved from fringe concern to operational risk.
This piece breaks down what's actually happening, where the real risk sits for recruiters, and how to design a hiring process that holds up.
The state of play: this isn't a one-off prank
The viral Reddit story is an anecdote. The pattern behind it is documented.
While these cases represent the most extreme end of the spectrum, the broader trend is equally concerning. Gartner projects that by 2028, one in four candidate profiles worldwide could be fake in some way. In a Gartner survey of 3,000 candidates, 6% admitted to some form of interview fraud: either sitting in as someone else or having someone else sit in for them.
Meanwhile, detecting fraud is becoming increasingly difficult: in a survey of U.S. hiring managers by Resume Genius, 76% said AI has made it tougher to spot impostor applicants, and 17% had already run into candidates altering video interviews with deepfake tools.
Several factors are driving this shift. Remote and hybrid hiring made the video call the default first impression. Generative AI has become more accessible, affordable, and capable. Face-swapping and voice-cloning tools that once required specialized equipment now run on a laptop.
The result is a hiring environment where the person on your résumé, the person in your video call, and the person who shows up on day one are no longer guaranteed to be the same individual.
Not All AI Use Is Candidate Fraud
Before discussing candidate fraud, it's important to draw a clear distinction: not all AI use is fraud.
Many candidates now use AI to refine résumés, organize experience, and practice interviews. That's preparation, not fraud. The real concern begins when AI is used to misrepresent identity or performance.
A useful way to think about it is through three categories:
-
AI-assisted preparation: using AI to polish materials or practice. Generally fine, and increasingly the norm.
-
Real-time AI answering: silently reading AI-generated responses during a live interview without disclosure. A gray-to-red zone that depends on the role and your stated rules.
-
Identity fraud: a deepfake interview, a voice clone, an AI agent, or a paid stand-in impersonating the actual candidate. This is the line that matters most, and the one that creates real exposure.
The goal isn't to ban AI. It's to establish clear boundaries and verify identity where the risks justify additional checks.
The Real Cost to Recruiting Teams
It's tempting to file fake candidates under "annoying." For TA leaders, the cost is more concrete than that.
It Undermines Hiring Quality
Every fake candidate who advances through the funnel consumes recruiter time, interview capacity, and hiring manager attention that should be spent on legitimate talent. The result is slower hiring, missed opportunities, and less confidence in hiring decisions.
The problem has become harder to detect in remote hiring environments. When sourcing, screening, interviewing, and onboarding happen entirely online, there are fewer natural checkpoints to verify that the same individual is present throughout the process.
Technologies such as face filters, voice cloning, AI-assisted answering, ghost-written assessments, and proxy interviewers all create opportunities for identity to shift unnoticed throughout the hiring process.
It Creates Security Risks
The real cost of candidate fraud isn't a wasted hire. It's a security risk. A mis-hire into an engineering, finance, data, or remote IT role isn't just a wasted seat; it can mean someone untrustworthy holding keys to source code, customer records, financial systems, and internal tools.The DOJ cases make the worst case unambiguous: impostors posing as remote workers reached more than 100 companies, and in some matters accessed sensitive data.
It Erodes Trust in Both Directions
Candidate fraud also taxes your relationship with every real applicant. To fight impersonation, teams add friction more verification, heavier screening, that friction reads as distrust to the genuine candidates who make up most of your pipeline. Strong candidates have options; when a process feels adversarial or over-automated, they disengage, decline, or talk about it publicly, and your employer brand takes the hit. For TA teams, trust is a measurable asset that shows up in conversion and offer-accept rates, so the rules you set should be clear and fair enough that honest candidates always know where they stand.
Five Ways to Reduce Candidate Fraud in Your Hiring Process
You don't fight this with a single deepfake detector. You fight it by designing a process where identity is confirmed at the right moments and where concerns are validated through evidence rather than intuition. Five moves matter most.
Publish your AI policy up front
State the rules where candidates actually see them: the job post, the interview invite, the assessment instructions. Spell out that using using AI for interview preparation is permitted; that no one may have another person or an AI agent attend an interview in their place; that reading from undisclosed AI-generated answers during a live interview isn't allowed; and, for technical assessments, exactly which tools are permitted. Clear expectations deter casual cheating and give you firm ground to stand on if something goes wrong.
Treat identity verification as a process
The strongest defense against a deepfake interview is making sure identity is consistent across the whole journey.
For higher-risk roles, build checkpoints at screening, technical interview, offer, and pre-boarding, and make the details line up: name, contact information, résumé, LinkedIn profile and GitHub history, government ID, work location, background screening, and onboarding documents should all tell the same story.
This mirrors the FBI's own recommendation to verify identity through interview, onboarding, and employment rather than once at the door. The goal of candidate verification is confirming that the same real person is present at every stage.
Rebuild your interview questions
Definition-based questions are easy to answer with the help of an AI copilot. Specific, follow-up questions rooted in a candidate's own work are much harder to fake in real time.
Instead of "what is X," ask:
● Why did you choose this architecture for the project listed on your résumé?
● What was the most difficult trade-off you faced?
● What would you change if you rebuilt the project today?
● Can you walk through the system architecture or debugging process live?
An impostor can recite an answer, but what they struggle to do is improvise convincingly about the texture, constraints, and second-guessing of real experience.
Use structured interviews
Verification should protect real candidates as much as it screens out fake ones. A shared red-flag checklist keeps interviewers calibrated:
-
unnatural movement or lighting around the face
-
audio that drifts out of sync
-
oddly delayed responses
-
repeated requests to restate questions
-
templated answers
-
an inability to explain résumé details
-
inconsistencies between stated location, education, or work history and what surfaces elsewhere.
The discipline that matters here is documenting what you observed, not eliminating someone on a hunch. Evidence can be reviewed; a feeling can't.
Have an escalation path ready
When an interviewer senses something is wrong, the worst move is to confront the candidate on the spot. Instead, have a clear escalation process: switch meeting tools, ask the candidate to briefly disable a virtual background, schedule a live work sample, bring a second interviewer into the next round, or route the case to HR for an identity-consistency check.
A defined escalation process ensures suspicious cases are reviewed and verified properly. It helps avoid awkward confrontations that can unfairly penalize legitimate candidates or alert fraudulent ones before an investigation is complete.
How an ATS Strengthens Candidate Verification
Most hiring teams view an ATS as a productivity tool. Increasingly, it also serves as a risk-control layer within the hiring process.
Candidate fraud often exploits the gaps between disconnected systems. Candidate data lives in one platform, interview feedback in another, and verification records in spreadsheets that are rarely revisited. A modern ATS close those gaps by centralizing sourcing data, résumés, communications, interview feedback, approvals, and offer records in a single candidate profile.
When every stage contributes to the same candidate profile, inconsistencies between interview conversations, application materials, and onboarding documents become much easier to identify.
This is where platforms like Greenhouse, Moka, and Ashby can help. As an ATS built for recruitment automation, it centralizes candidate information, supports structured interview questions and consistent feedback, interview summaries, and keeps compliance logs and audit trails of how each candidate moved through the funnel.
No hiring platform can guarantee detection of every fraudulent candidate. However, a centralized and traceable system gives hiring teams greater visibility, stronger process consistency, and the documentation needed to make decisions based on evidence rather than intuition.
Final Thought: Protect Your Process Without Alienating Candidates
As organizations strengthen fraud prevention, it's important to avoid overcorrecting. Anti-fraud measures that treat every applicant as a suspect can drive away the very candidates you want to hire. The goal is not to create more friction, but to build trust and consistency into the hiring process.
Candidates are often more receptive to verification than employers assume. Gartner found that 62% of candidates are more likely to apply when an organization requires in-person interviews, suggesting that reasonable verification measures can increase confidence rather than create resistance. When communicated clearly, verification can signal professionalism and fairness rather than suspicion.
Candidate fraud is unlikely to disappear, and both fraud tactics and detection methods will continue to evolve. What organizations can control is the integrity of their hiring process.
By building verification into the hiring journey, establishing transparent expectations, and using an ATS as part of a broader verification strategy, hiring teams can reduce fraud risk while protecting both hiring quality and candidate experience.
