The AI Hiring Arms Race Is Becoming an Infinite Doom Loop

It starts with an honest effort: a data scientist named Beggs sends out a job application, carefully tailored to the role. Then an automated system spits back a score -- 32 out of 100 -- with suggestions on how to reformat her resume and pack in more keywords. "If my job is to please the robots," she tells Wired, "then even if I don't like AI-generated content, maybe the machines do."

That single thought captures the logic driving the AI hiring arms race. Every week, tens of thousands of job seekers feed their resumes into ChatGPT, Claude, or Gemini and ask for rewrites optimized for Applicant Tracking Systems (ATS). Every week, tens of thousands of employers feed those same applications into their own AI tools to automatically rank, filter, and discard them. Both sides are using AI against each other. Both sides are getting more sophisticated. And according to the people who build the systems, nobody is winning.

"There's this tragic situation where each side has a problem. They're using AI to solve their own problem, but in ways that make the problem worse for the other side," Daniel Chait, CEO of ATS provider Greenhouse, told Wired. "The more it happens, the worse it gets."

The ATS mythology

At its most basic level, an ATS collects job applications and helps hiring teams track them. But the folklore around these systems has taken on a life of its own. An entire industry of resume-optimization services, keyword-matchers, and ATS-compatibility checkers exists to help candidates game systems they can barely see.

The reality is messier. Chait says no two ATS platforms work the same way. Some organizations use AI ranking heavily. Others -- including Toshiba and the remote-first company Doist -- insist that every application is reviewed by a human. "We have humans review every application," says Kim Jones, Toshiba's VP of HR. Automated ranking happens, she says, but it is far from universal.

This patchwork creates a paradox: because job seekers do not know whether an algorithm or a person will read their materials, many prepare for both. That means AI-polished applications written in a generic, keyword-stuffed style that satisfies neither a machine looking for authentic signals nor a person looking for genuine voice.

When AI screens AI

Doist ran an experiment that should give every job seeker pause. The company took several roles that had already been filled, fed the full applicant pools back through their ATS, and asked the system to produce a shortlist. The result? "In two instances that we tested, the person we hired wasn't in the short list," says Nadia Vatalidis, Doist's head of people. Their new hires -- people working out great after six months -- had been eliminated by the algorithm.

That finding echoes recent research from IEEE Spectrum and Fortune, both of which documented the growing trust gap between employers and candidates in AI-mediated hiring. A Clutch survey from July 2026 found that while 61% of job seekers said AI eased their search, 92% feared it would shrink the overall job market. The optimism is personal; the pessimism is systemic.

The candidate's countermove

James Jacobsen, a design professional who spent five months job-hunting, built his own antidote to the ATS black box. He used Claude to comb through job listings, analyze descriptions against his qualifications, and assign each one a score based on role type, seniority, and salary requirements. He built what is essentially the opposite of an employer's ATS -- a tracking system for the job seeker.

He reported that the amount of time he sank into searching was equivalent to a full-time role. Despite his AI-powered process, the results were mixed: a few interviews, no offers.

The infinite loop

The dynamic has a self-reinforcing quality that economists describe as a doom loop. Candidates submit more applications because they know fewer will be read. Employers deploy more aggressive AI filters because they receive more applications. Each side's escalation validates the other's. The system produces more volume, less signal, and growing distrust.

"The first time I can remember when both sides are unhappy," Chait says. "It's just not working."

Recruiters report receiving hundreds of nearly identical applications, each one clearly AI-generated. Hiring managers complain about candidates who pause mid-interview to type into ChatGPT. Candidates speak about the demoralizing experience of sending fifty applications and hearing nothing back -- black holes where their resumes disappear without so much as a rejection email.

What breaks the loop

There are glimmers of a way out. Jones of Toshiba notes that she almost never sees a cover letter anymore, which means applicants who include one stand out immediately. Chait advises job seekers to research companies actively rather than spray applications everywhere. Vatalidis suggests that the most effective candidates are those who treat the application as a communication between people, not between machines.

But these are individual coping strategies, not systemic fixes. As long as the fundamental incentives stay the same -- candidates needing to stand out in a flood, employers needing to manage volume -- the AI arms race will keep escalating. Each side will deploy better AI, which will be met by better counter-AI, driving the loop deeper.

The irony is impossible to miss: the technology that was supposed to make hiring more efficient has, in many ways, made it less human. The tools meant to connect talent with opportunity are instead building walls between them. Until someone finds a way to change the incentives, the doom loop keeps spinning.

The AI Hiring Arms Race Is Becoming an Infinite Doom Loop

Sources

Want to learn more?

Related Articles