How do I know if a candidate is using a deepfake on a video interview?

Watch for lag between audio and lip movement, flat lighting that never changes, and a candidate who avoids turning their head or holding an object up to the camera. Ask them to do both live. A real person can turn sideways and hold their ID next to their face in two seconds. A deepfake struggles with both. If something feels off, move the interview to a live in-person or proctored round before you submit them.

Will verification steps slow down my placement speed and hurt the candidate experience?

A little, and that is fine. Identity checks add about 3 to 5 minutes per candidate when you use the right tools. Real candidates do not mind proving they are real. The ones who ghost the moment you ask for a live ID check were going to burn you anyway. You lose a few minutes and save yourself a fraudulent placement that costs a client relationship.

Do I need to buy expensive software to stop fake candidates?

No. Start with process, not tools. A live camera check, a work-history call to a real reference, and a skills test the candidate takes on camera catch most fraud for free. Add paid identity verification once your volume justifies it. Most firms placing under 50 people a month can get 80 percent of the protection with zero new software.

The 2026 Verification Stack for Fake Candidates

AI-generated resumes and deepfake interviews just became the top hiring problem, ahead of talent shortage. Gartner says 1 in 4 candidate profiles will be fake by 2028. Here is the exact verification stack to build before your clients lose trust in you.

Fake candidates just passed talent shortage as the number one hiring challenge. That is not my opinion. It comes straight from GoodTime's 2026 Hiring Insights Report. For decades the hardest part of staffing was finding enough good people. Now the hardest part is figuring out which of your applicants are real humans.

Gartner projects that 1 in 4 candidate profiles will be fake by 2028. Read that again. One in four. AI-generated resumes, stolen identities, and deepfake video interviews are flooding your pipeline right now. If you place people for a living, this is your problem whether you want it or not.

Here is what makes it worse for staffing firms specifically. When a fake candidate slips through to a client, the client does not blame the candidate. They blame you. Every fraudulent submission chips away at the trust you spent years building. Lose enough trust and you lose the account.

What Fake Candidates Actually Look Like in 2026

This is not the sloppy fraud from five years ago. The tools got good. Here is what your recruiters are up against.

AI-written resumes that perfectly match your job description because the candidate fed your posting into ChatGPT and asked it to write a matching work history.
Stolen or borrowed identities where the person on the resume is real but the person interviewing is not.
Deepfake video interviews where a live face-swap runs during the call. The voice may be real. The face is not.
Proxy interviewers where a skilled ringer takes the technical screen and a different person shows up for the job.

The tech roles get hit hardest because they pay well and the work is remote. But no sector is safe. I have seen fraud in healthcare staffing, light industrial, and finance in the last year.

The Verification Stack That Actually Works

You do not fix this with one tool. You fix it with layers. Each layer catches what the last one missed. Here is the stack I build with staffing clients, in the order you should add it.

Layer 1: The Live Camera Check

Start every video screen with a 10-second identity moment. Ask the candidate to turn their head side to side and hold their photo ID up next to their face. Real people do this without blinking. Deepfakes lag, glitch, or the candidate suddenly has a reason they cannot do it.

This one step is free and catches more fraud than any software you can buy. Make it a required part of your first screen. No exceptions.

Layer 2: Resume Verification, Not Resume Reading

Stop treating the resume as fact. Treat it as a claim you have to prove. Pick two or three specifics from the work history and verify them on a live call. Ask about the team size, the tools they used, and a project that went sideways.

A real person answers with messy detail. A fake one gives clean, general answers because they are reading from a script the AI wrote. When you dig into specifics, the fraud falls apart fast.

Layer 3: On-Camera Skills Testing

Do not let candidates take skills tests alone at home. That is where the proxy takes over. Have them share their screen and solve a small problem live while you watch. For a developer, that is a short coding task. For an accountant, a quick spreadsheet exercise.

You are not testing whether they are a genius. You are confirming the person who interviewed is the person who can do the work.

Layer 4: Identity Verification Software

Once your volume climbs past 50 placements a month, add a paid identity check. These tools match a government ID against a live selfie and flag stolen documents. Names to look at include Persona, Clear, and iDenfy. Expect to pay a few dollars per check.

Do not start here. Software without process is a waste. Build the human layers first, then let software handle the volume you cannot cover by hand.

Layer 5: The Reference Call You Actually Make

Most firms skip references or take a list of buddies the candidate handed over. Instead, find one reference yourself through LinkedIn or the company directory. Call a real former manager, not a name off the candidate's list.

This closes the loop. If the work history is fabricated, the reference call is where it dies.

What This Means for Your Client Trust

Your clients are reading the same reports you are. They know fake candidates are surging. The staffing firms that will win the next two years are the ones who can say, with a straight face, that every person they submit has been verified as real.

That becomes a selling point. Put your verification process in your pitch deck. Tell the client exactly how you confirm identity, work history, and skills. Most of your competitors are still pretending this problem does not exist. You get to be the firm that solved it.

The cost of getting this wrong is not one bad placement. It is the account. A client who catches a fake candidate from you once will wonder how many others got through. That doubt ends relationships.

Do This Week

Pick one thing. Add the live camera check to every first-round video screen starting Monday. Write the exact script your recruiters will use, the head turn and the ID hold, and drop it into your interview template. It costs nothing and it catches the majority of deepfake attempts.

Once that is running, layer in the rest. You do not need to build the full stack overnight. You need to start before the next fake candidate reaches one of your clients.

Your Recruiters Are Drowning in Fake Candidates: The Verification Stack That Actually Works in 2026