AI Job-Search Arms Race Traps Candidates and Employers in Self-Defeating Loop
Applicants use AI tools to game applicant tracking systems while employers deploy AI screening to filter overwhelming application volumes, but evidence suggests the strategy often backfires.
What happened
Job seekers are using AI-powered resume optimization tools like Jobscan to tailor applications for automated applicant tracking systems (ATS), adjusting formatting, keyword density, and even stylistic choices like replacing "percent" with "%" based on algorithmic feedback. Employers simultaneously deploy AI screening within their ATS to rank candidates, believing only the top 10–20 percent will be reviewed. However, reporting from recruiters, HR managers, and company experiments reveals the premise is inconsistent: some organizations use AI ranking extensively, while others rely entirely on human review regardless of application volume. Doist, a small fully remote company, experimented with ATS ranking on previously filled roles and found that in two instances tested, candidates they hired did not appear in the AI-generated short lists. Daniel Chait, CEO of ATS provider Greenhouse, describes this as an "AI doom loop" where both sides use AI to solve their own problems—job seekers trying to pass filters, employers trying to manage volume—but in ways that worsen the underlying issues.
Context
The job market is currently characterized by scarce postings, declining trust between candidates and employers, and widespread application opacity. Job seekers invest substantial time with no feedback; employers report receiving hundreds of nearly identical applications, prompting them to automate screening. The proliferation of AI-driven optimization tools creates a perverse incentive structure where job seeker tools profit from the belief that ATS gaming is necessary, while the actual effectiveness of this strategy remains unproven and potentially counterproductive. The inconsistency in ATS adoption across organizations means applicants cannot reliably know whether algorithmic optimization will help or waste effort. This arms race also surfaces at the interview stage, where some candidates use AI to generate answers, which hiring managers can detect through telltale pauses and verbose responses. The core tension is that AI adoption by one side drives adoption by the other, amplifying the problem each side sought to solve.
What's disputed
Whether automated candidate ranking is actually standard practice in hiring is contested. Some organizations adamantly do not use AI screening and rely entirely on human review; others do use it. Greenhouse CEO Chait notes that "folklore" about ATS systems abounds, particularly from job seekers, and that no two ATS platforms work the same way or necessarily include AI at all. One recruiter demonstrated that strong interviews and an offer resulted despite poor Jobscan optimization scores, suggesting the tool's predictive value is limited.