Both sides brought machines to a human fight
Something is broken in hiring, and for once both sides agree. Jobseekers are using artificial intelligence to write applications at industrial scale; employers are using artificial intelligence to screen them at industrial scale. The result is not efficiency. It is an arms race — the BBC has asked whether an "AI war" is complicating job searches, and what the London non-profit People Like Us has bluntly called an "AI war" — in which the casualty is the thing hiring is supposed to discover: the actual human being.
The application used to be a front door. Now, too often, it is a dead end.
A woman looking over a job application form
The numbers behind the stalemate
The 2026 Greenhouse Future of Hiring Candidate Report — a survey of more than 2,800 jobseekers across the US, UK, Ireland, Germany and Australia — captures the deadlock in stark figures. More than three in four jobseekers (77 per cent) used AI in their search in the past year. Yet four in five (81 per cent) were rejected from a job in the past year without ever speaking to anyone at the company — human or AI. Nearly a third say this happens often or almost always, rising to 38 per cent of Gen Z in the US figures.
To fight through the noise, 71 per cent took extra steps to get noticed: seeking referrals (nearly half), paying for CV-writing services (27 per cent). And even landing an interview ends nothing: 96 per cent of candidates who interviewed had at least one complaint — most often that the employer simply ghosted them (45 per cent). As Greenhouse's chief people officer Sharawn Tipton puts it: "For the first time I can remember, neither side thinks hiring is working. Candidates send applications into a black hole, while employers sift through AI-polished CVs that all look the same."
The wider jobs picture is already under strain — US employers added just 29,000 jobs in September, and UK vacancies have fallen to a five-year low. The AI arms race is making a thin market feel impenetrable.
The cruellest irony: hiding the AI you used
Here is the catch-22 at the heart of the war. The same candidates who rely on AI are quietly rewriting their applications so employers cannot tell. Six in ten have changed — or considered changing — how they write so their application is not mistaken for AI-generated content.
Nearly two-thirds (62 per cent) have tried strategies to get around automated screening, from using AI to work out what the screening software is looking for (54 per cent) to — and this should alarm everyone — adding skills or keywords they do not actually possess (40 per cent). A third of those candidates hid text intended only for AI reviewers. Consider the absurdity: workers must now be fluent in two deceptions at once — one to pass the bot, one to pass as human.
A discrimination risk hiding in plain sight
There is a darker edge to all of this. A Censuswide survey of 2,000 UK workers, reported alongside the BBC's coverage of the issue, found that 75 per cent were unaware they have a legal right to challenge automated rejections. Worse, 70 per cent of ethnically diverse jobseekers said they had altered or concealed information on their CVs out of concern that recruitment software would screen them out — compared with 57 per cent of white candidates.
Automated systems trained on historical hiring data carry historical biases forward at machine speed — and the candidates with the most to lose are contorting their applications the most. This is the part of the "AI war" that no vendor's recruitment software advertises.
The employer's side of the black hole
Employers are not winning this war either. A TestGorilla survey of 1,428 US talent professionals in August 2026 found that 53 per cent believe all or most of the CVs and profiles they see are AI-written. But their number-one sourcing challenge is no longer spotting the AI — it is verifying that the skills listed are real (57 per cent).
When every application is polished by the same tools, the document stops being a signal. The CV, for a century the opening move of hiring, is depreciating like a currency in hyperinflation. Tipton again: "The application used to be a front door. Now, too often, it's a dead end." Employers who thought AI screening would save them time are instead drowning in plausible, interchangeable candidates. It is the same dynamic playing out in Barclays' AI-driven job cuts — automation reshaping work faster than the systems around it can adapt.
What actually works: talk to a human
The most useful data point comes from a smaller, harder survey. Career coach Jeremy Schifeling asked 236 jobseekers and 59 career centres what actually produced offers — and the answers cut through the hype. AI helps, but only as an assistant, never as the author. When AI wrote most or all of the application, the offer rate fell to 7 per cent, versus 12 per cent when it merely helped. Letting the machine write for you nearly halves your odds.
Volume strategies fail hardest of all: candidates firing off 51 or more applications a month had the lowest offer rate of anyone — 5 per cent. And jobseekers who had zero human conversations in a month were nearly three times as likely to quit their search altogether. Schifeling's one-line summary deserves framing: "AI gets you into the pile. People get you out of it." The algorithm can format your story. Only a person can vouch for it. More of our technology and opinion coverage explores how AI is reshaping work.
The rules of the new game
So where does this leave the jobseeker in 2026? First, treat AI as a spellchecker for your ambition, not its author — draft it yourself, then let the machine sharpen it. Second, spend the time you saved on the one channel the bots cannot fake: referrals, conversations, the warm introduction. The data is unambiguous that human contact is the differentiator.
Third, know your rights: the Information Commissioner's Office guidance says people subject to certain automated decisions can request a human review — a protection three-quarters of UK workers do not know exists. And the pressure on employers is building: People Like Us has launched a "Reject the Rejections" campaign demanding that companies audit their screening systems, publish the results, and submit to independent assessment. Its argument is that opacity is the real scandal — candidates deserve to know when a machine judged them.
My own view: arms races end the way they always end — not with a better weapon, but with an agreement. The employers who publish how their systems work, keep a human in the loop, and judge skills rather than prose will win the talent. The jobseekers who lead with their real voice will win the jobs. The machines can have the paperwork. The hiring should stay human.
