AI Is Turning Job Hunting Into a Self-Defeating Arms Race

AI is making the job market harder for both applicants and employers, as job seekers use AI tools to outsmart applicant-tracking systems while employers increasingly rely on those same systems to sort through floods of applications. The result, according to recruiters and hiring leaders, is a trust crisis that can trap both sides in a worsening feedback loop.

That dynamic is forcing candidates to optimize résumés for algorithms that may not even be doing the screening, while companies complain about walls of nearly identical applications and turn to software for faster triage. The practical outcome is fewer human connections, more automated guessing, and a hiring process that many participants say feels broken.

For job seeker Jodi Beggs, the experience became absurdly literal: small formatting tweaks, such as using a middle initial consistently or replacing the word “percent” with a symbol, changed how an online résumé-scoring system judged her materials. Her takeaway was blunt: if the machines are judging the application, then candidates may feel pressure to write for the machines, even when they do not trust the process.

Why the job search feels stuck in an AI loop

The modern applicant-tracking system, or ATS, was built to help employers collect, review, and manage applications. In practice, it has become a catch-all term for software that may include résumé parsing, keyword matching, candidate ranking, interview scheduling, onboarding workflows, and, in some products, automated screening tools powered by AI.

That complexity is part of the problem. Job seekers often assume there is a single hidden machine that decides whether a résumé lives or dies. In reality, ATS products vary widely by vendor, customer settings, and the features a company actually pays to use. Some organizations use AI ranking; others do not. Some rely on software only for administration and keep human review at the center of hiring.

Still, the belief that software is making the first cut has become powerful enough to shape candidate behavior. People now write and re-write résumés for invisible systems, use job-match tools to score themselves, and strip their experience down to keywords they think an algorithm wants to see.

Daniel Chait, CEO of ATS vendor Greenhouse, said the market is full of myths about how these systems work. His broader warning was that the job-seeker side and the employer side are each trying to solve their own problems with AI, but often in ways that intensify the same frustrations.

That is the loop: candidates use AI to beat AI; employers use AI to manage AI-shaped volume; and both sides end up less satisfied than before.

What Jobscan-style tools promise — and what they miss

Tools such as Jobscan are built on a simple promise: compare a résumé against a job posting, identify missing keywords, and increase the odds of passing an ATS screen. For candidates overwhelmed by unemployment, that promise can sound like survival advice.

But the promise rests on an assumption that is often incomplete. In many companies, automated ranking is only one piece of a larger human process. In others, there is no AI ranking at all. That means a high score from a résumé optimizer can be irrelevant, or at least much less important than the tool implies.

The business model also raises questions. Subscription-based résumé helpers often charge monthly fees, which means they are not especially motivated to tell users to stop searching and apply less. The incentive is to keep candidates engaged, refining materials and trying again.

That is one reason hiring experts caution against over-relying on these services. They may help with formatting, clarity, and keyword alignment, but they cannot guarantee an interview, much less a job offer.

What the ATS actually does

An ATS is, at minimum, a database and workflow tool for hiring. It stores applications, tracks candidates through interview stages, and helps hiring teams coordinate decisions. Some systems also manage recruiting emails, background steps, or onboarding.

Depending on the company, an ATS may also support the following:

• Résumé parsing and search
• Keyword matching against job descriptions
• Automated candidate ranking
• Interview scheduling
• Screening questionnaires or AI interviews
• Offer and onboarding workflows

The key point is that “ATS” is not one universal machine. It is a category of products, and the actual experience of candidates depends on configuration, company policy, and human judgment.

How employers are responding to application overload

Employers are not using AI in a vacuum. Many hiring teams say they are buried under large pools of candidates, including applications that appear nearly identical after being polished by generative tools. When the pile becomes unmanageable, software starts to look like the only scalable answer.

That is especially true in a weak hiring environment, where postings are scarce and applicants may send out dozens or hundreds of applications for each role. Candidates report hearing nothing back for long stretches. Employers, meanwhile, complain that the signal-to-noise ratio keeps getting worse.

Some states are beginning to examine job-market fraud and ghost jobs, which only compounds the suspicion. If a candidate cannot tell whether a role is real, and an employer cannot tell whether a résumé is authentic, trust collapses quickly.

Once that trust erodes, the temptation to automate grows stronger. AI becomes a defensive tool rather than a strategic one.

How different companies are handling hiring

Interviews with recruiters, HR executives, and small-business owners show that there is no single industry standard. Some teams use AI-powered ranking. Others insist that every application is reviewed by a human. The deciding factor appears to be company culture more than company size.

Toshiba is one of the firms that still emphasizes human review. Kim Jones, vice president of human resources, says every application is looked at by people. She is not concerned if applicants use AI to clean up their materials, but she argues that a stronger résumé alone will not matter if the candidate does not meet the core requirements for the role.

Jones also pointed to a newer issue: AI-assisted interviewing. In her experience, some candidates appear to be leaning on generative tools during live interviews, with noticeable pauses followed by polished but overly long answers.

Jones said that applicants sometimes seem to be typing during interviews before delivering a more elaborate response, suggesting they may be consulting AI in real time.

Doist, a small remote-first company that hires internationally, has also experimented with AI-based ranking. But the company found that the results did not reliably identify the people it ultimately hired.

Nadia Vatalidis, the company’s head of people, said the team ran tests on previously filled roles by feeding job descriptions and saved application materials into AI shortlisting systems. In at least two cases, the eventual hires were not included in the short list generated by the system. The overlap between interview candidates and top-ranked candidates was only partial, and the people who later proved successful were not always the ones the software favored.

Why the “AI doom loop” keeps getting worse

The phrase “AI doom loop” captures a simple but troubling cycle. Employers face too many applications and turn to automation. Candidates suspect automation is deciding their fate and respond by using AI to reverse-engineer the system. Employers then encounter even more polished, more uniform, and more machine-optimized applications, which pushes them further toward software filters.

That feedback loop does not just reduce efficiency. It changes behavior on both sides of the market.

1. Job seekers spend more time tailoring each application.
2. Employers spend more time sorting through flooded pipelines.
3. Both sides rely more heavily on AI tools.
4. Neither side gains much confidence in the process.

Chait argues that this dynamic is making the system less useful, not more. The more each side depends on AI to solve a trust problem, the more trust erodes.

How applicants are adapting

Some candidates are now treating AI as a job-search assistant rather than a résumé rewrite machine. That can mean using language models to compare openings, organize applications, track follow-ups, and rank opportunities by fit.

James Jacobsen, a design professional who has been job hunting for months, took that approach. He started by using tools such as Claude and ChatGPT to polish application materials, but found that alone did not significantly improve results. So he widened the use of AI to manage the search itself.

He created a system that scanned job listings, summarized descriptions, tracked notes, and scored opportunities based on factors such as seniority, type of work, pay, and work arrangement. He used the model to remind him why he had previously rejected a role if it appeared again later.

In effect, he built a personal version of the system employers use to track candidates.

He also used AI to review and restructure his portfolio, which led to a major overhaul of his presentation materials. Even with that effort, the process still did not translate into an offer.

What job seekers can do instead of spray-and-pray applying

Hiring leaders interviewed for the story emphasized that candidates may be better served by focusing on quality over volume. Instead of sending the same application to dozens of postings, they recommend targeted outreach, company research, and human networking.

That advice matters because the job market is not just a technical filtering problem. It is also a relationship problem. When applications arrive as interchangeable documents, the person reading them may have no reason to remember one candidate over another.

For that reason, even small differentiators can matter. Jones said she rarely sees cover letters anymore, which means the ones that do arrive stand out. Whether or not a cover letter is read closely, it can still signal intent and effort.

Networking remains another underused tactic. A warm introduction, a referral, or a conversation with someone inside a company can bypass the cold mechanics of a blind application process and make a candidate more memorable.

Practical lessons from recruiters

• Customize the application, but do not obsess over algorithm myths.
• Research companies before applying to understand fit and culture.
• Use AI as an organizer, not just a keyword-stuffing device.
• Write a cover letter when appropriate, especially if the role seems competitive.
• Prioritize networking and referrals where possible.

The biggest shift, according to Chait, is philosophical: job seekers may need to accept that the old mass-application strategy is becoming less effective. More volume does not necessarily produce more interviews, especially if everyone else is using the same tools to flood the same openings.

How much should candidates trust ATS advice?

Candidates should treat ATS optimization as one input, not a guarantee. Résumés still need to be clear, concise, and relevant, but overfitting to a guessed-at algorithm can be a waste of time. In many cases, the better strategy is to make the application readable by both people and software.

That means avoiding clutter, using standard section headings, keeping job titles and dates consistent, and aligning experience with the role without distorting the truth. It may also mean resisting the urge to turn every resume into a keyword puzzle.

Job seekers can improve their odds by focusing on three things: clarity, relevance, and evidence. A readable résumé that shows measurable impact is useful whether a recruiter scans it manually or software ingests it first.

What this means for the labor market now

The deeper lesson from this hiring stalemate is not simply that AI is changing recruiting. It is that AI is being inserted into a part of the economy that already suffered from opacity, inefficiency, and power imbalance.

For job seekers, the search has become more time-consuming and less legible. For employers, candidate volumes can be overwhelming and authenticity is harder to assess. Both sides feel manipulated by a system that promises efficiency but often delivers more work.

The danger is that the market starts optimizing for the appearance of precision rather than actual better matches. A résumé scoring 95 instead of 82 may not mean much if the model is misaligned with the role, the recruiter’s goals, or the company’s hiring culture.

In that sense, the job market has become a test of judgment not just for people but for the tools they use. And right now, many of those tools appear to be amplifying the very problems they were supposed to solve.

Bottom line for candidates and employers

The core message from recruiters, HR leaders, and candidates is straightforward: the hiring process is increasingly being shaped by AI, but not necessarily improved by it. Job seekers are trying to out-engineer the system, employers are trying to tame application overload, and neither side is fully winning.

That is why the current environment feels like an arms race with no finish line. If AI is used only to game a broken process, the process becomes harder to trust. If the process cannot be trusted, both sides lean harder on AI. The loop feeds itself.

For now, the most durable strategy may still be the least glamorous one: target roles carefully, connect with people, submit thoughtful applications, and remember that the humans still matter, even when the software gets the first look.

Chait’s final advice to job seekers was simple: the problem is bigger than any one candidate, and the system itself needs fixing.