The ghost candidate problem: what the research reveals (and what you can do about it)

 

We recently facilitated a conversation with Laura Schroeder (COO of LHH Recruitment Solutions) and Brad Miller (Global Client Director at LinkedIn) about one of the thorniest hiring problems of 2026: ghost candidates. Hundreds of talent professionals tuned in with questions. How do you spot AI fraud? How do you make your application stand out? How do you keep the human element in hiring when everything feels automated?

On spotting fake candidates and resumes

Can you really detect AI-generated resumes?

The short answer: it's harder than most recruiters think. About 75% of US job seekers now use AI to polish their applications, which means a polished resume no longer signals effort or fit. It signals access to AI.

Three things that matter more than resume polish:

1. Intent signals. Did they apply to one well-matched role or spray applications across your entire career track? Did they reference something specific about your team, your product, or the problem you're solving?
2. Behavior clues. Did they complete optional steps? Did they follow through on interviews, or did they ghost? Did their story hold up across resume, LinkedIn, and conversation?
3. Early verification. Adding an identity check or document verification early in the process keeps deepfakes and synthetic candidates out before you invest recruiter time.

The real insight from our survey: 99% of HR leaders worry about candidate misrepresentation, and only 18% are very confident their process can catch it. The gap between widespread concern and low confidence is a vulnerability.

What about fake AI avatars in interviews?

In the past 12 months, 40% of surveyed teams saw candidates use AI in ways that raised concerns during interviews, 37% suspected someone interviewing wasn't the real applicant, and 33% suspected deepfake video interviews.

What works: bring candidates in early for a live interaction (video call, phone screen, casual conversation) before the formal interview. Live human contact is hard to fake at scale. It's also when you can ask clarifying questions that draw out whether they actually did what's on their resume.

For candidates: how to get past the noise and land an interview

How do I get my application actually reviewed by a human?

The application volume right now is massive: LinkedIn processes 11,000 applications per minute, most of which never reach a human.

Here's what can get you noticed:

• Apply to roles you're really qualified for. It may seem obvious, but one targeted application to a well-matched role beats 50 spray applications.

• Reference something specific. Mention a recent product launch, a team member you know, a problem they're solving. Show you did your homework.

• Lean into what AI can't fake easily. Your projects, your work samples, your thinking on their specific challenges. If you can, send unrequested work that speaks to their problem.

• Reach out directly to someone on the team with a personalized message, not a mass LinkedIn InMail. A referral is still the strongest way to get an interview.

• Complete optional steps. A portfolio, a video intro, a sample project. These are friction points that filter out low-intent candidates and show you're serious.

What about job boards and ghost postings?

There are a lot of fake job postings out there. How do you avoid applying to them?

• Check if the company is hiring. Go to their careers page. Do they have this role listed there?
• Look at the company's LinkedIn. Are they actively posting about hiring? Do recent updates mention this team?
• Check job board posting dates. Does the posting feel active or is it months old and still circulating?
• Verify with a direct connection at the company, if you have one.

For HR teams: volume, process, and the resource crunch

How do we handle volume without adding more screening layers?

47% of surveyed HR teams report increased recruiter workload, 46% have added screening or verification steps, and 42% report wasted interview time for hiring managers.

Adding more screening doesn't always fix it. What does:

1. Filter for intent at the front. Knockout questions (Does this role match your background? Have you worked in this industry?) and high-intent signals (referrals, specific knowledge of the role, unrequested work samples) thin the pile before it reaches your team.
2. Verify identity early. One ID check or document verification early in the process keeps fraud out before you invest time. Do this before you spend recruiter resources.
3. Reduce the application friction in the right way. If one-click applications are bringing in 242 applications per opening, make the threshold slightly higher. Ask a targeted knockout question. Make candidates choose between roles instead of applying to everything.

How do we protect the human moments?

64% of HR leaders say screening and administrative work leaves them too little time for real conversations with candidates. Those conversations are what actually moves hires forward.

Protect human time by:

• Offloading volume screening. Use a recruitment partner, a sourcing team, or structured screening questions to thin the pile before it reaches your hiring manager.
• Shifting from "catch fraud" to "spot fit." Instead of hunting for red flags, ask questions designed to reveal real capability and intent. Can they explain why this role fits their trajectory? Can they speak to your customers or recent work unprompted?
• Reframing the technical assessment. Instead of testing whether they used an AI tool during an interview, test whether they can think through your actual problem. Bring them in to brainstorm. Watch them work. That's harder to fake.

What if we just don't have the resources to add verification steps?

Partner with someone who does. One of the most direct recommendations from our research: hand screening to a recruitment partner who's built for high volume. They do the sorting, verifying, and vetting, then deliver HR a short list of real, interested people.

This isn't outsourcing decisions. It's outsourcing friction so your team can focus on what matters: talking to actual candidates and making hiring decisions.

On the bigger picture

Is there relief coming for job applicants?

Maybe, but not yet. Someday software may surface real candidates at the push of a button. Until then, both sides are stuck playing this game.

For job seekers: the best move is being real. Do work that shows your thinking. Apply to roles you actually want. Engage directly with people at companies you're interested in. Be findable through referrals and your network.

For recruiters: the best move is redesigning your process around signal, not volume. Get a smaller, real, verified pool of genuinely interested people in front of a human who can actually evaluate them.

What about career pivoting?

Yes, it's harder, and the research backs this up. When AI has made it possible to apply at scale, candidates often spray applications across industries instead of building a coherent story. Hiring teams see a resume that claims to pivot from everything.

If you're pivoting: build your narrative. Show progression and intention. Invest in projects or work that demonstrates capability in the new space. This is hard to fake and it signals you're serious about the move, not just applying to everything.

The real answer

Our panelists landed on something important: the answer to the AI problem isn't necessarily more AI.

It's process. It's reframing what you're screening for. It's human empathy. It's asking your team where they're confident in your hiring process and where they're not, then fixing those gaps.

Laura Schroeder, COO of LHH Recruitment Solutions, framed it this way: instead of asking "how do we detect AI," ask "does our process reveal the human's real capabilities, skills, and fit?"

Reframe that question in your own process and you're halfway there.

---

 

Want the full research? Download our paper "When Effortless Applications Made Hiring Harder" to see the full survey data (500 HR and talent acquisition leaders, conducted in 2026) and detailed recommendations.