Gartner predicts that by 2028, one in four candidate profiles worldwide will be fake. That forecast comes from a Gartner press release dated July 31, 2025. It is a prediction about where hiring is heading, not a measured rate for today. The same release reports something that is measured: in a 2Q25 Gartner survey of 3,000 job candidates, 6% admitted to interview fraud, either posing as someone else or having someone else pose as them.
If you hire remote engineers, that's the number to keep in mind. Here is what it means for a buyer, and what we do about it at Ruzora.
Key Takeaways
- Gartner's figure is a forecast for 2028 about candidate profiles. Treat it as a direction of travel, not a statistic about your current pipeline.
- AI makes two kinds of fraud cheaper: the wrong person showing up, and the right person borrowing skills they don't have.
- Ruzora engineers pass two separate evaluations, a graded coding challenge and an AI interview answered on camera, and they need a passing score on both before they reach our bench.
- You still meet the engineer yourself before anyone starts work, and a 60-day replacement guarantee covers technical performance from their first working day.
What "Fake" Means When Candidates Have AI
Most of the coverage blurs two different problems together.
Identity fraud is the wrong person. Someone applies under a borrowed or invented identity, or a stronger engineer sits the interview for them. Gartner's 6% figure is about exactly this: posing as someone else, or having someone else pose as you.
Capability fraud is the right person with borrowed skills. The candidate is real, but an assistant writes the code or feeds the answers during the assessment. The profile looks senior, and the work in month one doesn't.
AI lowers the cost of both. A polished CV takes minutes to generate. A plausible answer to a textbook interview question is one prompt away. Gartner's Jamie Kohn warned in the same release that candidate fraud creates cybersecurity risks "far more serious than making a bad hire." An engineer gets access to your code, your cloud, and often your customers' data in week one.
None of this means remote candidates are suspect by default. Most are exactly who they say they are. It means a vetting process has to be built so that borrowed identities and borrowed skills don't survive it.
How We Vet Engineers Against AI Fraud
We are deliberately vague about how our checks work. People who cheat read blogs too, and a check that is fully described is a check that can be rehearsed. So this section describes what each step guarantees for you, not how it does it.
| What we do | What it means for you |
|---|---|
| A graded coding challenge in the engineer's stack | The code is scored on its own, as its own evaluation |
| An AI interview, with every answer recorded on camera | The engineer explains their work out loud, in their own words. No camera, no interview |
| Both evaluations are required | A strong score on one never makes up for a weak score on the other |
| Blind profiles first, identity after the service agreement | You judge the work before the name, then meet the person |
| Your own live interview before anyone starts | The last check is yours, with the engineer in front of you |
Two separate evaluations, both required
Every engineer on our bench has passed a graded coding challenge and an AI interview. They are scored separately and we never average them. An engineer needs a passing score on both before they reach our bench. That matters for fraud: a candidate who leans on outside help in one setting has to hold up in a different setting, with different questions, on camera.
On camera, in their own words
In the AI interview, the engineer answers each question out loud on webcam. That is where borrowed skill tends to show. Someone who can produce an answer is not always someone who can explain the trade-offs behind it, follow up on their own reasoning, and talk about real work they've done.
You meet them before they start
Our flow puts you in control of the final step. You see blind profiles with their scores first. When you pick who you want to meet, Roberto sets up a quick call, you sign the service agreement (MSA), and then we schedule interviews with the engineers' full identity. Run your own live interview. Ask them to walk you through something they built. Nothing replaces that conversation, and we don't pretend it does.
A guarantee that starts on day one of work
If an engineer isn't working out on documented technical grounds in their first 60 days, counted from their first working day, we replace them at no extra cost. That's what our contract says, and it's why the vetting has to hold up.
The Honest Counterpoint
No vetting process stops every scheme, and I won't claim ours does. Fraud adapts, and any check that works today will be tested tomorrow. What a layered process does is raise the cost of faking it at every step, so the people who make it through are overwhelmingly the people they say they are.
That is also why you should keep running your own checks on anyone you hire, including engineers we send. How to identify fake candidates in remote hiring covers the FBI's warning signs and a stage-by-stage process, and deepfake interview detection covers what to do on a video call.
Check a Candidate You Already Have
If you're weighing a candidate right now, you don't have to take anyone's word for it. Our free candidate check sends you a link to forward to them, and you get a scored technical report once they finish. We only email your candidate if you give us their address and ask us to. Then, in your own interview, ask them to talk you through how they solved it.
If you'd rather start from engineers who have already passed both evaluations, tell Sol what you're hiring for and get a vetted shortlist within 72 hours.
Frequently Asked Questions
Will 1 in 4 candidate profiles really be fake by 2028?
That's Gartner's prediction, published July 31, 2025: by 2028, one in four candidate profiles worldwide will be fake. It's a forecast, not a measurement. The measured number from the same release is that 6% of 3,000 surveyed candidates admitted to interview fraud.
How do you catch fake or AI-assisted candidates?
Every Ruzora engineer passes two separate evaluations before they reach our bench: a graded coding challenge and an AI interview answered on camera, in their own words. A candidate needs a passing score on both. You then interview the engineer yourself before they start. We don't publish how our checks work, because that would make them easier to beat.
Do your engineers use AI coding tools on the job?
Many do, and most clients expect it. The point of vetting is to confirm the engineer understands and can explain the work themselves. See how to vet developers who use AI coding tools.
The Bottom Line
Gartner expects one in four candidate profiles to be fake by 2028. You don't need to panic about that, and you shouldn't ignore it. Hire through a process where every engineer passes two separate evaluations, answers on camera, and still meets you before day one. Start with the free candidate check, or ask Sol for a shortlist.
Roberto Espinoza is CEO of Ruzora, which helps US startups hire pre-vetted senior LATAM engineers, with a vetted shortlist in 72 hours. See available engineers.
