Between 2022 and 2025, Greenhouse tracked more than 640 million applications across over 6,000 companies. Applications per job rose from 116 to 244. Applications per recruiter rose from 146 to 746. Recruiters per organization fell from 10.43 to 4.62.
Those four numbers describe one event: the cost of submitting an application fell to approximately zero, while the cost of reading one did not move at all. A candidate reporting nine months and a hundred applications is not describing a personal failure. They are describing the output of that arithmetic.
The Volume Is Real & It Is Mostly The Same People
LinkedIn now processes roughly 11,000 application submissions per minute, a 45% increase year over year, and attributes much of the rise to generative AI. What changed is not the number of people looking for work. A candidate who once sent ten tailored applications a week can now send several hundred, auto-written and auto-submitted overnight for about twenty dollars in tooling.
The Willo Hiring Trends Report 2026 found 77% of hiring teams regularly encounter AI-generated or AI-assisted applications, against 53% in early 2024. When every cover letter has been polished by the same handful of models, the cover letter stops carrying any signal at all.
Both Sides Optimized Locally & The System Degraded
Greenhouse CEO Daniel Chait calls the pattern an AI doom loop. Candidates use AI to apply more, employers use AI to filter more, rejected candidates respond by applying still more. Every individual decision in that sequence is rational. Chait’s summary of where it lands is worth repeating: this is the first time when really both sides have been unhappy.
His platform currently hosts around 175,000 live jobs averaging 254 applicants each. For an HR function, the important implication is that no amount of screening capacity solves this, because the screening capacity is what the other side is optimizing against.
Duration Is Where The Damage Concentrates
The headline unemployment picture obscures what is happening at the tail. BLS data puts the median duration of unemployment at around 11 weeks in mid-2026 while the mean sits near 24, and more than one in five unemployed workers has been searching for 27 weeks or longer. A United Way NCA survey of 1,000 recent and current job seekers found an average search running 28.5 weeks, about 62.6 applications, and more than 46 hours spent on applications alone. The averages describe a long middle. The people writing the posts your executives read are in the tail, and the tail is growing.
What Actually Reduces The Noise
Changes that move the ratio rather than adding screening capacity:
- Publish a real salary range, which is the single strongest self-selection filter available
- Ask one or two role-specific written questions at application, since these cost an auto-applier more than they cost a real candidate
- Set a recruiter SLA with a stage-level clock and monitor anyone not contacted inside it
- Close or refresh postings on a fixed cadence, because stale requisitions are most of what candidates call ghost jobs
- Report applications-to-qualified-candidate as a funnel metric, not applications received
- Send an actual rejection at every stage past the first screen, automated where necessary
Friction Is Being Reintroduced On Purpose
The industry consensus reversed quickly. Ophir Samson, who leads voice AI at Greenhouse, told WIRED that recruiters who spent years asking for the easiest possible apply flow now explicitly request friction, after seeing two thousand applicants arrive in twenty-four hours for a single role.
LinkedIn rolled out an underqualified-applicant warning in August 2026, notifying users when their profile does not match a role’s stated requirements. The design question facing every talent team is not whether to add friction but where to put it, since friction at the application stage filters differently than friction at the screening stage, and the second kind falls hardest on candidates without paid tooling.
Ghosting Has Become A Measured Risk
Criteria’s 2026 Candidate Experience Report found 53% of job seekers were ghosted by an employer in the past year, a three-year peak, with post-interview ghosting reported by a larger share still. Harvard career advisor Gorick Ng described the mood to Axios as having moved past ordinary job search burnout into systemic cynicism.
The behavior is now reciprocal, with roughly a quarter of job seekers reporting that they have ghosted an employer mid-process. A pipeline built on candidates who have learned not to invest in any single application produces exactly the no-shows and late-stage dropouts that recruiting teams are currently reporting as a separate problem.
Say What Your Rules Are
One practical move sits within any HR team’s control this quarter. Writing in Forbes Human Resources Council, Dr. Timothy Giardino argued that candidates should not have to guess an employer’s position on AI use, and proposed zoning it explicitly, with resume clarity, translation and interview preparation treated as acceptable and fabricated experience treated as not.
Publishing that guidance costs nothing and does two useful things. It gives honest candidates a reason to slow down and tailor, and it gives your team defensible ground when someone is screened out. The alternative is a process where both parties are running undisclosed automation against each other, and then wondering why the hire took nine months on both sides of the table.
