AI Job Market Creates ‘Infinite Doom Loop’ as Both Sides Race to Optimize

The American job market has entered what experts call an AI doom loop — a self-reinforcing cycle where both employers and job seekers deploy artificial intelligence to solve their respective problems, only to make matters worse for everyone involved.

Job postings are scarce in the current low-fire, low-hire economy. Scams and ghost jobs have become so prevalent that several states are considering new legislation to combat the problem. Both sides report a breakdown of trust: candidates invest enormous time into applications with no response, while employers say they receive hundreds of nearly identical resumes.

Chait described the dynamic plainly: more AI use begets more AI use, to no one’s benefit. The more it happens, the worse it gets.

The divide over AI adoption in hiring appears driven by philosophy and culture rather than company size. Kim Jones, vice president of human resources at Toshiba, insists every application is reviewed by humans. She noted that while applicants using AI to polish materials is fine, it does not help them get through automated applicant tracking systems. The real problem, she said, shows up during interviews — the pause, the typing sounds, then a verbose answer that clearly came from a machine.

Doist, a small fully remote company that hires internationally, ran its own experiment. When head of people Nadia Vatalidis fed past applicant materials into an AI ranking system, two out of two tested cases showed that the people they had actually hired — employees who were performing well after six months — did not appear on the AI’s short list.

James Jacobsen tried a different approach. A design professional who began his job search five months ago, he used Claude and ChatGPT to streamline his entire application pipeline — combbing listings, analyzing job descriptions, logging opportunities, and building a scoring system based on seniority, salary requirements, and work format preferences. When he rejected a position, he trained his AI to remember why, so reposted roles would trigger an automatic reminder. He spent less time searching and found higher-quality positions. He received a few responses but no offers.

As the cycle continues, both sides remain trapped in a system where optimization tools designed to create advantage instead produce uniformity, noise, and disillusionment.