AI broke the job application, and companies are scrambling to fix it

By the numbers: the average corporate job opening now draws somewhere between 242 and 254 applications. Do the math, and any single applicant's odds of landing that specific role come out to roughly 0.4%.

It gets more extreme at the entry level. In the UK alone, more than 1.2 million applications were submitted last year for fewer than 17,000 graduate roles.

• 11,000: job applications submitted on LinkedIn every minute.
• 242-254: average applications per job opening.
• 0.4%: one applicant's rough odds of landing any single role they apply to.
• 1.2 million: UK graduate-role applications last year, for under 17,000 openings.

Why it matters: job seekers can now buy AI tools, some for as little as $20, that automatically apply to every matching job posting they can find, no manual effort required beyond setup.

That turns job hunting into a numbers game rather than a targeted search. If applying to 400 jobs costs almost nothing in time or effort, why would anyone apply to just five, even if most of those 400 applications are a poor fit?

In real life it's like everyone at a crowded restaurant deciding to order every single dish on the menu, just in case one of them turns out to be good, and the kitchen has to somehow serve all of it at once anyway.

The catch: not as much as they used to. Only 37% of employers now see resumes and credentials as reliable indicators of actual talent, and 41% say they're actively moving away from resume-first hiring altogether.

That's a real, measurable shift in how hiring works, not just employer frustration venting. When a resume can be AI-generated to match any job description almost perfectly, it stops being useful as a filter.

Greenhouse CEO Daniel Chait has a name for the resulting spiral: the "AI doom loop." His description is blunt: "Everyone's using their own AI to solve their own problem, but it's making the whole system worse."

The loop works like this: job seekers get ignored, so they apply to even more jobs using AI. Employers get flooded, so they use AI to filter faster. Neither side's AI actually fixes the underlying mismatch, it just makes both sides move faster toward the same frustrating outcome.

Who's affected: Chait put it simply: "This is the first time when really both sides have been unhappy. The market just isn't working for either side." Job seekers pile applications into what feels like a black hole, and recruiters drown in submissions that took the applicant seconds to generate.

What's next: companies are turning to methods that are genuinely harder for AI to mass-produce. Skills tests, work samples, employee referrals, and AI-run interviews are all becoming more common as filters that require real, individual effort from each candidate.

Greenhouse itself has tried two specific fixes. A feature called "My Dream Job" lets candidates flag one priority role per month, and those flagged applications get hired at roughly 5 times the standard rate, nearly 500,000 have been submitted so far. Separately, Greenhouse acquired Ezra AI Labs to offer AI voice interviews to every applicant, reducing the incentive to mass-apply since every submission now gets a real interaction.

Background: the tension between applicants and hiring software isn't new, AI just sped it up dramatically. Applicant tracking systems, the software companies use to manage incoming resumes, date back to the late 1960s and 1970s, originally as simple digital filing systems.

Over the following decades, these systems evolved to scan resumes for specific keywords, automatically advancing or rejecting candidates based on how closely their resume's wording matched a job description. That created its own well-known workaround industry: job seekers learned to stuff resumes with the right keywords to get past the filter, whether or not those keywords reflected genuine experience.

Why it matters: AI just made both sides of that keyword arms race dramatically faster and easier to scale. Where a job seeker once had to manually tailor keywords for each application, AI now does it instantly, for hundreds of applications at once. Where an employer once needed a human to skim resumes for obvious mismatches, AI now filters thousands of submissions in the time it used to take to review a handful.

Background: the flood of AI-generated applications is landing on top of an already messy problem: not every job posting represents an actual open role. Multiple studies published in early 2026 estimate that around 47% of online job listings qualify as "ghost jobs," postings the company isn't actively trying to fill.

The reasons vary. Some companies keep listings live to build a pipeline of resumes for roles they expect to open eventually. Others use open postings to project growth to investors, or to convince overworked existing staff that relief is on the way. In a survey of over 1,600 hiring managers, 62% admitted to posting fake jobs specifically to make current employees feel replaceable.

Why it matters: that context makes the AI-application flood look even worse from a job seeker's side. People are often burning AI-assisted effort applying to roles that were never going to result in a hire in the first place, adding a second layer of wasted motion on top of the first.

Who's affected: ironically, the advice that works best right now is the least AI-scalable kind. Recruiters consistently point to networking, personal referrals, and skills-based demonstrations as more effective than a polished resume alone, precisely because those are the things AI tools can't mass-produce on someone's behalf.

That's a genuinely harder path for job seekers without existing professional networks. It also means a lot of quality job openings never get posted publicly at all, filled instead through referrals before the flood of public applications ever begins.

• If you're job hunting, a resume alone likely isn't enough anymore. Expect skills tests, work samples, or AI-run interviews at more companies.
• Mass-applying with AI tools may actually hurt you. With average odds already near 0.4% per application, spreading effort thin rarely beats a few genuinely tailored applications.
• Networking and referrals matter more than ever, precisely because they're the one part of hiring AI can't easily automate on your behalf.
• If you're hiring, resume screening alone is losing reliability fast. Consider skills-based filters before wading through hundreds of similar-looking applications.