If you have applied for a job this year and heard nothing, the newest data on hiring suggests two things can be true at once: the listing may have been real when you applied, and it may still be sitting there — unfilled, unclosed, collecting CVs — months later.
A report computed on 23 September by job-search tool Unlisted counted 607,050 postings open on employers’ own career sites, read directly from fifteen applicant tracking systems rather than scraped from job boards. Of those, 28.3 percent had been up for more than 90 days. That is 163,057 listings — and 94,106, or 16.3 percent, had been open more than 180 days.
🔍 THE BOTTOM LINE
The report measures staleness, not intent: it never claims any single posting is fake, and it notes that some roles are evergreen and some employers hire in volume all year. But combined with separate survey data showing hiring managers spending more time than ever filtering AI-written applications, the numbers describe a labour market where both sides of the desk are now drowning each other out.
What the numbers say
The median open posting is 36 days old. Age varies hugely by industry: hospitality postings sit the longest — 43.9 percent over 90 days, median 65 days — while healthcare turns over fastest at 19.7 percent. Engineering (32.3 percent) and skilled trades (29.5 percent) sit mid-table. Country-by-country, the United States (26.2 percent over 90 days from 334,681 postings) and the UK (26.8 percent) both run near the global average; Australia is lower at 21.5 percent.
The closures side of the ledger cuts the other way. Of 88,895 postings removed in the report’s 30-day window, 14.6 percent came down within a week — postings that vanish in days, Unlisted notes, were real. Only 4.2 percent of closed postings were reposted within 30 days as the same role at the same employer, which is the closest thing in the data to a “ghost” signature. The report is careful about its own limits: it has only been watching for 29 days, so closure-time figures are lower bounds, and it cannot say whether an employer that keeps a listing up is still interviewing.
The AI layer underneath
The staleness data lands in a market already tangled by generative AI on the applicant side. A Robert Half survey published in London in April found 67 percent of UK hiring managers said screening for AI-generated applications had increased time-to-hire, and Matt Weston, the firm’s Senior Managing Director for the UK and Ireland, said generative AI had made it “considerably harder to evaluate CVs at face value.” A separate March survey by the same firm found 67 percent of US HR leaders reported slower hiring because of AI-written application volume, with one in five reporting delays beyond two weeks.
Indeed’s CEO Hisayuki Idekoba described the loop plainly in September: applications per job have more than doubled since 2022 while recruiters per organisation fell 56 percent. AI on both sides has added volume without adding signal.
The counter-arguments worth reading
Reactions to the report, on Hacker News and elsewhere, include a serious defence of long-open listings. Recruiters argue a single posting for ten engineering hires stays open by design; specialty roles can take months even when the hiring is genuine. One comment made the sharpest counter-point: where postings stay up forever, the honest explanation is often not a fake job but a company that can’t keep people — brutal conditions or pay that will not clear the market. Hospitality’s 43.9 percent staleness rate, the highest of any category, is consistent with either reading.
For New Zealand job seekers the practical takeaway is the same either way. SEEK’s August report shows NZ job ads up 9.7 percent year on year — the 21st consecutive month of growth — with hospitality, tourism and retail hiring ahead of the summer season. Demand exists; the noise around it is the problem. Checking whether a listing is current on the employer’s own site, favouring roles posted in the last month, and treating months-old listings with suspicion costs an applicant nothing.
What job seekers can do with this
The report’s own framing is the most useful tool: it publishes posting ages by category and by applicant tracking system, free to quote and reproduce. A listing’s age is now public information for 95 percent of postings — the date the employer published is stamped on most boards. Age is not proof of anything, but it is a free filter, and in a market where 62 percent of job seekers now use AI tools to apply, the scarce resource is a human being reading your application.
FAQ
What is a ghost job? A listing that stays visible after a company has stopped actively hiring for it — sometimes a filled role that was never taken down, sometimes a posting used to collect CVs with no opening behind it. The Unlisted report counts postings open more than 90 days and avoids calling any single one fake.
Are ghost jobs increasing because of AI? The report does not attribute staleness to AI. What has changed is the surrounding market: AI-written applications have flooded employers — 67 percent of UK hiring managers told Robert Half the screening burden had slowed hiring — which lengthens pipelines and gives employers less incentive to close listings quickly.
How do I check if a job posting is stale? Look for the publish date on the employer’s own careers site; the report found it is visible on 95 percent of postings. Postings older than 90 days are in the 28.3 percent band; hospitality and education postings sit longest, healthcare turns over fastest.
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
