The AI Doom Loop in Hiring: What 412% More Applications Means for Your Interview

Daniel Chait, the CEO of Greenhouse, spent the last week talking to every major business outlet about something breaking in real time on his platform. His subject: an AI-driven spiral that has made the hiring process worse for everyone simultaneously — candidates and employers alike.

The number he keeps returning to: applications per recruiter have increased 412% since 2022. The average job posting on Greenhouse now draws 254 applicants. Some of those applicants are paying $20 for tools that automatically blast their profile to every open position on the platform, regardless of fit, skills, or interest.

"This is the first time when really both sides have been unhappy," Chait told Fortune on July 27. "Everyone's using their own AI to solve their own problem, but it's making the whole system worse."

That's the doom loop. Job seekers use AI to apply to more jobs, faster, with less friction. Recruiters use AI to filter the flood. Rejected candidates apply to even more jobs. Recruiters build higher AI barriers. Volume spirals. Neither side gets what they need.

What it produces, on your end, is a market where sending applications feels like sending email into a void.

What the Data Actually Shows

Chait's observation isn't anecdotal. Greenhouse hosts roughly 175,000 live job postings at any given time, and the platform's 2026 benchmarks are public. Those postings collect an average of 254 applications each. The average recruiter is processing 412% more applications per year than in 2022 — for the same number of open roles.

The composition of those applications has also shifted. AI-assisted resume writers produce documents that are keyword-dense and professionally formatted in a way that looks similar at scale. When every resume in a 254-candidate pool is polished, the signal that used to separate candidates at the application stage largely disappears. There's no longer a meaningful quality gap between a carelessly assembled application and a carefully crafted one, because AI has compressed the distribution.

This is the mechanism of the doom loop. The tool that was supposed to make each individual candidate more competitive has made the overall pool more homogenous. Each individual candidate is now harder to distinguish. The candidates who paid for mass-apply tools didn't gain an advantage — they joined everyone else in applying to jobs they won't hear back from.

What Employers Did About It

The employer response has come in two forms.

The first is defensive: more AI filtering. If AI-generated applications are flooding the top of the funnel, you deploy AI to manage the funnel. Greenhouse's own research found that 63% of active job seekers have now been screened by an AI at some point in the hiring process — up 13 percentage points in a single six-month period. Algorithmic resume ranking, AI-conducted phone screens, automated scoring before any human recruiter gets involved. The filtering infrastructure on the employer side has grown in direct proportion to the volume arriving from the candidate side.

The second response is more structural: redesigning what the interview itself is for.

Greenhouse launched a feature called "My Dream Job," free to job seekers on the platform, that lets candidates flag exactly one role per month as their top priority. Not one per company. One per month, across all 175,000 jobs on the platform. The constraint is the point. Scarcity creates credibility. When a candidate is forced to pick one role per month as their target, the signal they're sending to that employer — this is genuinely the job I want — is distinguishable from the noise of bulk applications.

The early results: candidates who used the "Dream Job" flag got hired at roughly five times the rate of other applicants, and landed roles in an average of 20.5 days compared to 35 to 50 days for candidates who didn't use it.

That's not a marginal difference. And it points precisely to where hiring is actually happening in 2026: employers are finding ways to identify candidates who have a specific, demonstrated reason to want this particular job — because those candidates stand out against a field where most applications carry no signal about intent at all.

What This Does to the Interview

The downstream effect on the interview itself is specific.

When 254 people apply for a role and only a few will ever speak to a recruiter, the candidates who get through represent a pool that has already been filtered for unusual intent or fit. But getting to the interview is only the beginning of a changed environment.

Research tracking more than 19,000 live interviews found that 38.5% of candidates showed signs of AI assistance during screening rounds — tools feeding answers through hidden overlays, transcription-based prompting, or generated scripts. The number hit 48% for technical roles. Employers have taken note, and the interview structure has shifted in response.

Companies are moving weight toward later rounds, toward live and in-person conversations, and toward follow-up questions designed to probe whether a candidate can extend and defend the answers they gave. The "walk me through your reasoning" and "how would your approach change if X were different" questions are not there to trip candidates up — they're there because those are exactly the questions AI-assisted answers struggle to handle under pressure.

This shift has one significant implication for anyone who lands an interview in 2026: the bar has risen. The interview is now structured specifically to separate surface-level performance from underlying capability. That's not an accident. It's the employer's direct response to a doom loop where the interview is the only stage left that AI hasn't made indistinguishable.

What to Actually Do With This

Recognize the interview for what it is. In a market where 254 people applied for this role, getting a conversation is a statistically unusual outcome. The preparation that's proportional to that rarity is more than a resume review — it means deeply understanding the company's actual situation, the specific people you're meeting, and the problem this role is meant to solve. Treat every interview like the hard-won, limited opportunity it is.

Prepare for depth, not coverage. The questions that derail candidates in 2026 interviews aren't the unexpected ones — they're the follow-ups on the questions candidates answered confidently. If you're going to talk about a project, know it thoroughly: the numbers, the decisions, what didn't work, your specific contribution. Answers that hold up through three follow-up questions require ownership you can't fake and can't generate on the spot.

Make your intent specific and audible. The "Dream Job" data reveals something broader than a platform feature: intent is the differentiating signal when everything else has been compressed by AI. In the interview, the candidates who advance are the ones who articulate a clear, specific reason why this job at this company is what they're actually trying to do — not a generic "I'm excited about the opportunity," but a real argument grounded in the company's situation and the candidate's relevant background.

Build thorough preparation before you sit down. Everything available to help you understand the role, the company, the interviewers, and the likely questions belongs in the preparation window before the call, not on a browser tab you're reaching for mid-interview. Meeting Copilot's interview assistant is built for this: load your resume, the job description, and research on the role so your preparation is organized and accessible throughout the conversation rather than scattered and out of reach when you need it.

The System Is Broken on Both Sides

The doom loop Chait described isn't a failure of job seekers or a failure of employers. It's a systemic outcome where both sides rationally adopted AI tools to solve their own problems and collectively made the situation worse.

The aggregate picture is stark: applications per recruiter up 412%, an average of 254 applicants per posting, and a labor market with 7.6 million open jobs that produced only 5.2 million hires in the same month — according to the Bureau of Labor Statistics' June 2026 report. Openings and hiring have decoupled in part because AI volume has degraded the signal quality at the top of the funnel.

The fix isn't more AI at the application stage. The "My Dream Job" data makes that clear: what cut through wasn't better resume optimization. It was demonstrated intent and the scarcity that made that intent credible.

For the practical reality of job searching right now: the application stage is where volume goes to die, on both sides. The interview is where hiring actually happens — and where intent, preparation, and real conversational capability are the deciding factor.

Getting into that conversation is the genuinely hard part in 2026. Once you're there, it's the part you can control.

---

Sources: Fortune: Greenhouse CEO on the AI Doom Loop · HR Dive: Hiring in an AI Doom Loop · Greenhouse: Dream Job Launch · Metaintro: Employers Rebuilding Interviews · Newsweek: 22% Use AI During Interviews · BLS JOLTS June 2026 · Greenhouse: 63% Interviewed by AI