The AI Job-Hunt Doom Loop: Both Sides Are Using Bots and Nobody Is Winning

Key points

• Jobscan, one of the most popular résumé-optimisation tools, costs $30 to $50 per month and may not reflect how any specific employer actually screens candidates.
• An experiment at software company Doist found that AI ranking tools failed to short-list the candidates who were eventually hired and thriving six months later.
• Daniel Chait, CEO of hiring platform Greenhouse, says automated AI screening and AI-assisted applications have created a feedback loop that makes hiring worse for both sides.
• Human resources VP Kim Jones says Toshiba has humans review every single application, and a cover letter now genuinely makes a candidate stand out.
• Some job seekers are spending the equivalent of a full-time work week on applications and still receiving no response.

Jodi Beggs is a data scientist looking for work. She recently discovered that her two-page résumé lost her points in an automated scoring tool, that using her middle initial on one document but skipping it on another also counted against her, and that swapping the word "percent" for the % symbol pushed her score up. None of that has anything to do with whether she can do the job.

Beggs was testing one of a growing category of apps that compare your application materials to a job listing and tell you how well they match. The best known is Jobscan. The underlying idea is that many employers now use an ATS, an applicant tracking system, which is software that collects and sorts job applications, to automatically rank candidates before any human lays eyes on them. The theory goes that only the top 10 to 20 percent of applicants make it through. Everyone else disappears.

Is that actually how hiring works?

Sometimes, but not always. Daniel Chait, CEO of Greenhouse, an ATS company, says there is a lot of folklore here. Some organisations do use automated ranking. Others have humans read every application from the start. No two systems work the same way, and the features a company uses depend on what they have paid for and what their culture values.

Kim Jones, vice president of human resources at Toshiba, is direct about it: "We have humans review every application." She does not object to candidates using AI to polish their materials, but warns it will not magically get them through. What actually matters to her team are job requirements, salary expectations, and a handful of practical factors.

Doist, a small fully remote company that attracts international applicants, ran a telling test. They took roles that had already been filled and fed the job descriptions plus all the original applications into an AI ranking system. Their question: would the people they had actually hired show up on the AI's short list? In two separate tests, the eventual hire did not make the cut at all.

What does the doom loop actually look like?

Chait describes a spiral that is hard to escape. Employers receive hundreds of near-identical applications, many polished by AI tools, so they lean on automated ranking to sort them quickly. That makes job seekers more anxious and more likely to use AI résumé optimisers to game the system. Which produces even more look-alike applications. Which pushes employers further toward automation.

"More AI use begets more AI use, to no one's benefit," Chait says. "The more it's happening, the worse it gets."

Designer James Jacobsen found a more creative way to use AI. Rather than just tuning his résumé, he built a personal job-search tracker using Claude and ChatGPT, the two most widely used AI chatbot tools. He had them scan listings, score opportunities against his own criteria (seniority, salary, remote versus office), and log reasons he had rejected a role so he would not waste time if it reappeared. He spent less time searching and found better matches. He still has not landed an offer, but he is working smarter, not just harder.

What should job seekers actually do?

Skip the spray-and-pray approach, where you blast the same application to hundreds of jobs and hope for the best. Chait's advice is to research specific companies you would genuinely like to work for, even ones that are not household names.

Jones at Toshiba adds something that surprised me: she almost never sees a cover letter anymore. A tailored one now makes a real impression, precisely because it is rare. Networking, old-fashioned as it sounds, remains underrated.

And if the silence is getting to you, remember Chait's message to candidates: "It's not you; it's the system, and it stinks."

Tool or approach	Cost	Likely benefit	Catch
Jobscan résumé scorer	$30 to $50 per month	Keyword alignment tips	May not reflect your actual employer's system
AI chatbot (Claude, ChatGPT)	Free tier available	Drafting, critiquing, tracking	Output can sound generic
Tailored cover letter	Free	Stands out because it is rare	Takes more time per application
Networking	Free	Direct access to real humans	Slower, but Chait says it works

Common questions

Do all employers use AI to automatically filter applications?

No. As first reported by Wired AI, the reality varies widely. Some companies, including Toshiba, have humans read every application. Whether AI screening happens depends on the employer's culture and which software features they have switched on.

Is it worth paying for a résumé optimiser like Jobscan?

Probably not as your main strategy. One recruiter showed she scored poorly on Jobscan yet landed a dozen interviews and a job offer. At $30 to $50 a month, the service has little financial reason to tell you to stop subscribing.

Can AI help with job searching in any genuinely useful way?

Yes, particularly for organisation. Using a chatbot to track applications, score listings against your own priorities, and draft tailored cover letters is more effective than using it simply to stuff your résumé with keywords.