Ask almost any job seeker and you’ll hear the same refrain: AI has turned applying for work into a series of hoops to jump through.
Candidates use ChatGPT to write their résumés and cover letters; employers turn tools of machine learning on the deluge of submissions; everyone ends up feeling worse. In a Reuters/Ipsos poll, 71 percent of respondents said they fear AI will take away too many jobs for good — a concern that crops up more and more among the dispiriting messages sent by frustrated job applicants.
That anxiety isn’t occurring in a vacuum. The labor market is cooling around the edges. Long-term unemployment is at a post-pandemic high, according to The Washington Post. New job growth in recent months is approximately zero, analysts at Citi have said, according to NBC News. And Stanford economists recently discovered that generative AI is already correlated with fewer new job postings for software developers — one of the fields many expected to benefit most from the technology.
OpenAI’s big hiring play
Into this ferment, enters OpenAI which is developing an AI-driven hiring platform to launch in 2026, industry publications report. The goal: Aid employers in finding talent that can use AI productively — and get candidates to show off skills remaking a resume, not just dumping keywords into an internet search engine. Think LinkedIn crossed with an assessment engine — only with OpenAI models at the wheel, both for matching and verification.
OpenAI said it has been talking with large employers, including big retailers, consulting firms and state agencies in order to map what “AI-literate” work looks like in practice. The company’s online learning hub, OpenAI Academy, will issue certificates that show up on candidate profiles. For employers, the pitch is a reduced signal-to-noise ratio; for applicants, evidence of competency beyond an ATS (applicant tracking system)-optimized resume.
Fidji Simo, the head of applications at OpenAI, has written that AI will create opportunity much more than any other technology before it. It’s a bold promise to be making at a time when not even leaders of AI are willing to take such a bullish stance—tae, the CEO of Anthropic has warned that automation could eradicate a “substantial proportion” of menial white-collar work by 2030. That tension establishes the stakes for OpenAI’s gambit: Can more AI really untangle the AI-induced bottleneck?
The arms race that broke recruiting
The hiring funnel that has emerged now is, in some ways, a perfect storm.
Low-friction application tools and AI-generated submissions have inflated volume. Recruiters depend on automatic filters to survive, and the risk that qualified candidates are being screened out by rigid rules or brittle keyword checks is rising. Now applicants respond by gaming the system even harder, sometimes using the same instruments that were swamping the system all along. The end result is more like people as pieces of work than, say, performance art for algorithms.
AI patches are already promised by platforms. LinkedIn is adding generative tools for resumes and recruiter outreach. Enterprise suites such as Workday and Eightfold AI also use machine-learning algorithms to rank candidates. Niche services such as Hiring. cafe and Sonara automate the job search for seekers. But candidate experience scores have not taken off, and time-to-hire continues to be a stubborn stat in many industries, HR associations report.
If OpenAI would like to avoid this fate, it must address three issues: noisy signals, opaque decisions and weak trust. Employers not only need more candidates; they want validated skills. Candidates don’t want to wonder why they didn’t make the cut; instead, they need feedback that encourages them to improve. Both sides need confidence that the system is not quietly perpetuating bias.
What would actually help
Skills-first matching beats resume-first filtering. Assessments embedded in authentic tasks — a prompt to engineer an email from the customer, a data-cleaning notebook, or a sales scenario — can and do evidence competence far more effectively than buzzwords. If OpenAI’s platform values these signals more than keyword scans, then the gulf between potential and pedigree may be slightly narrower.
Transparency is also nonnegotiable. Applicants would be able to see which factors mattered, whether or not a human being reviewed their profile and what they could do to improve a future application. Regulatory bodies already are headed in this direction: New York City’s Local Law 144 mandates bias audits for automated employment tools; the EEOC has put out guidance on AI in hiring; the EU’s AI Act lists employment as a “high-risk” use that calls for close monitoring. Each platform that is incorporating explainability and auditing from day one will be in a better position than those that attempt to retrofit compliance at a late stage.
Finally, interoperability matters. If skills proofs and certificates are locked inside one ecosystem, then that platform becomes another walled garden.” Employers will, of course, continue to live on LinkedIn, Indeed, Workday and their own ATS. An open standard for candidates to take their verified achievements across systems would help keep the market honest — and hold applicants back from endlessly retaking the same tests.
A shaking market, and a short window
None of that changes the macro picture overnight. Analysts point to cooling job-creation momentum, and studies indicate AI is already reshaping demand in white-collar roles — reducing postings for some jobs but increasing the need for workers who can effectively wield these tools. The World Economic Forum has estimated that we will see a net loss of jobs over the next years as automation outstrips new role creation – at the same time as demand for AI-related skills skyrockets.
That’s why the timing matters. If OpenAI’s landing pad is just another automated gate, it will compound the malaise. If it really damps noise, exposes decisions and boosts evidence-based skill, then perhaps markets will feel fairer — without going to the lengths of pretending AI can make demand appear out of nowhere in an economy absent a bread-and-circus politics. The bar is high, but the need is great. Applicants don’t need more AI in hiring; they simply need better AI, applied transparently, to find actual work. Now OpenAI has to show that that’s more than marketing.