90% of employers rejected real candidates over suspected "AI-written" résumés, and half now run AI interviews themselves
Recruiters are so wary of AI-written résumés that they are screening out real people too. In a Greenhouse survey of US employers, 90% said they had rejected genuine candidates because an application looked "AI-generated." Yet half of the companies now use voice AI for first-round interviews, and 40% of those say it surfaced strong candidates they would otherwise have missed. The takeaway for HR products: résumé text is losing its weight, while a short conversation is becoming more valuable.
AI-processed from ZDNet AI; edited by Hamidun News
Picture a recruiter with hundreds of identically polished résumés and cover letters on their desk. They are clearly written with ChatGPT, and telling a strong candidate from a weak one by the text is nearly impossible. What do you do? Many simply discard anything that "smells" of AI. A new survey from hiring platform Greenhouse shows the scale of the problem: 90% of US employers believe they have rejected real people because their applications looked fake or AI-generated. In the same breath, the second figure: half of companies themselves use AI to conduct first voice interviews with candidates. The market distrusts AI-written text while hiring an AI interviewer.
Suspicion of "too perfect" résumés
Greenhouse surveyed 1,347 recruiters and hiring managers in the US, UK, Germany, Australia and Ireland. The survey ran from late September to mid-October.
The company itself adds a caveat. The data is self-reported: it reflects what employers believe, not a measured fact. Greenhouse also does not know whether companies run applications through AI-text detectors or whether recruiters judge "by eye." So "90%" does not mean "90% of good candidates were lost" but rather "90% of employers think they lost someone good."
The response to the problem is quite practical. 86% of respondents changed their hiring process because of AI, and 57% of them added another interview with a live person. It resembles checking a signature: when people stopped trusting the paper, they pick up the phone and ask in person.
There is also irony here. Automated résumé filters reject most applications, candidates tailor their formatting to satisfy them, and as a result the texts look even more "machine-made" and provoke even more suspicion.
Voice AI as the first interviewer
The second part of the story looks almost the opposite. According to Greenhouse, half of employers already use AI interviews to hear from more candidates at an early stage. 40% of them say they found strong people they would otherwise have passed over.
The company offers an example. One HR specialist ran 200 already-rejected candidates through Greenhouse's voice tool, and the system recommended re-examining 19. All five finalists came from this group, and one person was hired. Software developer Ryan Quinlan looked like a "maybe" on paper, but in the conversation he mentioned skills he had not listed on his résumé, which helped him advance.
It is important to remember that Greenhouse sells this tool itself, so the figures come from an interested party. The company says the AI does not decide whom to advance or reject, and does not rank candidates. Every result is reviewed by a recruiter who is accountable for the outcome and can explain how it was reached. Another detail is telling: of all the tasks employers are willing to hand to an AI agent, selecting candidates for the next stage ranked last, at just 7%.
What candidates themselves think
According to an earlier Greenhouse report, only 7% of US job seekers would refuse an AI interview, and 70% of those who had one say it helped them show their skills. But such numbers are easy to read as desperation. The same data shows that 81% of job seekers were rejected over the past year without talking to anyone, either a human or an AI. Against that backdrop, 15 minutes of conversation with an algorithm beats a résumé lost in a stack.
Greenhouse representative Ofir Samson asks people not to reduce everything to desperation: candidates set conditions. Nearly half (49%) want to know whether a human will review their answers, and 40% want to understand the criteria by which they are assessed. The company adds that candidates can opt out of AI interviews and that about 90% of participants rate the experience as excellent, but this is the company's own claim.
Practical takeaway
For developers and product managers of HR technology, the lesson is this. First, detecting "AI text" is unreliable, and the cost of error is high because you lose real people. Second, dialogue works as a check that is harder to fake than a résumé. Third, transparency becomes part of the product: candidates want to know who reviews the results and on what criteria. Fourth, a human making the final decision is becoming a market requirement, and California already has a law against dismissals based on AI assessment without human sign-off.
Open questions remain. It is unknown how many good candidates AI screeners filter out themselves, and what exactly "improved candidate experience" means in employers' answers. Meanwhile, 85% of companies that saved time through AI say they reinvested it in candidates: more communication with the shortlist (45%) and more early interviews (40%).
- Nov 21, 2026Mozilla с GPT-6 Cyber нашла и исправила в 8 раз больше багов Firefox за год
- Nov 21, 2026Perplexity returned risky shops in 15% of shopping answers, ChatGPT in 8%
- Nov 21, 2026OpenAI Names 4 Reasons GPT-6 Astra Hides Its Reasoning from Monitors
- Nov 21, 2026Australia's Central Bank: AI Hits Young Workers Aged 15–24 First in IT and Customer Support
Key Questions Answered
What are the key details on Suspicion of "too perfect" résumés?
Greenhouse surveyed 1,347 recruiters and hiring managers in the US, UK, Germany, Australia and Ireland. The survey ran from late September to mid-October. The company itself adds a caveat. The data is self-reported: it reflects what employers believe, not a measured fact.
What are the key details on Voice AI as the first interviewer?
The second part of the story looks almost the opposite. According to Greenhouse, half of employers already use AI interviews to hear from more candidates at an early stage. 40% of them say they found strong people they would otherwise have passed over. The company offers an example. One HR specialist ran 200 already-rejected candidates through Greenhouse's voice tool, and the system recommended re-examining 19. All five finalists came from this group, and one person was hired.
What candidates themselves think?
According to an earlier Greenhouse report, only 7% of US job seekers would refuse an AI interview, and 70% of those who had one say it helped them show their skills. But such numbers are easy to read as desperation. The same data shows that 81% of job seekers were rejected over the past year without talking to anyone, either a human or an AI. Against that backdrop, 15 minutes of conversation with an algorithm beats a résumé lost in a stack.
Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews
Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.