Ask any recruiter you know how the last quarter went. The honest answer involves a number that wasn't true two years ago and a feeling that wasn't there last summer.
LinkedIn applications hit 14,200 per minute in February 2026, up 58% from 2024. The number of U.S. applicants per open role has doubled since 2022. The inbound channel that used to deliver qualified applicants now delivers volume, and most of that volume is indistinguishable from the rest on the surface. AI-generated résumés can be produced and submitted at near-zero marginal cost, and they're getting better every quarter.
Three responses have emerged. None of them are wrong. None of them are bulletproof. And none of them answers the question that's actually breaking the funnel.
1. Buckle Down and Grind Through It
More recruiters, more hours, more careful manual review. Effort produces output, but it doesn't scale. The volume has grown faster than headcount, and the manual reviewer cannot reliably tell an AI-generated application from a real one once both clear the basic keyword checks. Greenhouse CEO Daniel Chait calls the resulting state an "AI doom loop."
2. Deploy AI to Parse What's Real
Fight AI with AI. ATS platforms have added scoring layers. Recruiters are using LLM-powered tools to identify suspicious phrasing, detect generated content, and cross-reference profiles. Among Talent Acquisition professionals surveyed by LinkedIn, 93% said they planned to grow their AI use in 2026. But AI screening filters for quality signals like writing style, profile consistency, and plausibility. It does not verify identity. A perfectly written, plausible application from a fabricated person passes AI screening cleanly.
3. Abandon Inbound and Go Outbound
Stop waiting for applications altogether. Recruiters are reaching out directly on LinkedIn to candidates with strong profiles, encouraging them to apply rather than waiting for the inbound flood. The pipeline is smaller. The candidates are more qualified. This is the cleanest response of the three, and it's working for the teams that have made the shift.
But it has its own blind spot, and it's the same one the other two haven't solved.
What Is Outbound Sourcing Actually Verifying?
Picture how the outbound sequence actually unfolds. A recruiter sees a strong GitHub profile linked from a senior engineer's LinkedIn. They send an InMail. They get a thoughtful reply within the day. They schedule a video call. The candidate shows up on camera, answers questions fluently, walks through the projects on their resume. Three rounds later, an offer is signed. Two weeks after that, the new hire logs in for the first time.
At every step in that sequence, the recruiter trusted that the same person was on the other end. The LinkedIn profile, the GitHub, the InMail reply, the face on the video, the signature on the offer, the laptop on Day 1. Six handoffs. Zero verification.
Every one of those trust points used to be reasonable. Faking a LinkedIn profile required time and effort. Faking a GitHub history required real technical work. Faking a video interview was not commercially feasible. That changed quickly. AI-generated profile photos are indistinguishable from real ones. GitHub histories can be fabricated or purchased. Deepfake video filters are commercially available and improving every quarter.
The most documented example of this playing out at scale is also the most uncomfortable. In 2025, federal prosecutors revealed that a single facilitator in Arizona had helped North Korean operatives land jobs at 309 U.S. companies, including Fortune 500 names, over three years, generating $17.1 million in fraudulent salary payments. The operatives used stolen U.S. citizen identities, AI-generated profile photos, deepfake video filters during interviews, and a residential "laptop farm" to make their work appear domestic. CrowdStrike documented a 220% year-over-year increase in this exact pattern. The FBI has issued public advisories specifically calling on employers to add identity verification at every stage of hiring.
Outbound has reduced volume and improved candidate quality. It has not verified identity. It moved the identity problem from the inbound channel to the outbound one, and the recruiter is now trusting a chain of assumptions that AI has systematically devalued.
What Are All Three Approaches Actually Missing?
The hiring funnel was designed to answer one question. Is this candidate qualified?
That question used to include identity implicitly. When candidates walked into an office for an interview, identity verification happened in the handshake. The person who got hired was visibly the same person who interviewed, and visibly the same person who showed up on Day 1. The hiring process verified qualification, and physical presence verified identity at no extra cost.
Remote hiring removed the physical step without replacing it. The funnel now answers the qualification question through screens, calls, and documents, none of which were designed to confirm that the human on the other end is who they claim to be.
The recruiter buckling down, the recruiter running AI screening, and the recruiter sourcing on LinkedIn are all working on the same underlying problem from different angles. The problem isn't a recruiting problem. It's an identity assurance problem the hiring process has never had a tool to address.
What Are HR Leaders Starting to Do About It?
The leaders getting ahead of this aren't choosing between the three approaches. They're adding the layer all three have been missing.
That layer is identity verification as a discrete step in the hiring process, separate from sourcing, separate from screening, separate from the background check. Not "did this candidate pass the assessment." Not "does this identity have a clean record." But "is the human being who accepted the offer the same human who appears on Day 1." Confirmed biometrically against a government-issued ID, before access is provisioned.
This isn't a fourth filter in the funnel. It's a verification question that travels with the candidate regardless of how they entered the process, and every recruiter benefits from it, whether they're sourcing on LinkedIn, running AI screening, or grinding through manual review.
If you want to know whether you have the layer in place, a question worth asking your team is: For the last ten people we hired, how many can we prove were the same person on Day 1 as the person we interviewed? If the honest answer is anything less than ten, that's the work.
Why Is HR the Right Function to Lead This?
The identity chain in any organization starts in hiring. Once a hire is made, every downstream system trusts that the person provisioned is the person hired. The CISO can harden the perimeter, but they cannot retroactively verify an identity that was never confirmed at the source.
HR is the only function with the authority and the workflow to verify the human at the source, before that identity propagates through every downstream system. Not as a security mandate handed down from the CISO, but as a strategic capability that makes the hiring process defensible against the threat patterns that have emerged in the last two years. The HR leader who builds an identity-verified hiring pipeline is the one who can answer the FBI advisory, the CISO's question, and the CEO's inevitable follow-up.
Recruiting answers the qualification question. Background checks answer the history question. Identity verification answers the human question. All three need an answer, and only one of them is currently absent from most hiring pipelines. HR is the function with the authority and the workflow to add it.
