AI Hiring in 2026: Half of Job Seekers Were Rejected Without a Word

“It told me I was AI.”

That’s a line one U.S. job seeker wrote into our open-ended survey question about the strangest experience they’d had with hiring software. They weren’t a bot. They were a person applying for a job, evaluated by an algorithm that decided they didn’t seem human enough.

A few years ago, that would have been an outlier. It isn’t anymore.

In April 2026, Enhancv surveyed 1,066 U.S. job seekers about the receiving end of AI hiring. Half of them (50.5%) had been rejected at least once in the past year without a single word from a human. Of that group, 63.8% said they believed a machine made the call. Only 9.7% of the full sample said an employer had ever clearly told them AI was involved.

Everyone else was guessing. Most still are.

The data below is what that silence looks like at scale. One in three candidates has walked away from a job rather than sit through a one-way AI interview. Nearly half are using AI of their own to handle the ones they didn’t walk away from. The rest are guessing—at the rules, at the rubric, at what the machine wants from them. Most never had a way to know.

Let’s take a closer look at the data.

Are job rejections going silent in 2026?

Hiring used to come with a sentence. A “We went with another candidate.” A phone call. A form letter. Even an automated email signed with a recruiter’s name. The wording varied, but it was confirmation that a person, somewhere, had read the application.

Most candidates aren’t getting that anymore.

More than half of the respondents have been rejected without a single word from a human in the past year. Add the “Maybe” bucket and the share rises to 57.1%.

It doesn’t fall on everyone equally. Younger candidates are roughly twice as likely as those over 55 to be on the receiving end of it. This suggests that AI gating concentrates at the entry and early-career tiers where applicant volume—and resume parsing—runs hottest.

Here’s the silent rejection rate by age group:

• 25–34 (Late Gen Z + Millennial): 62.2%
• 18–24 (Gen Z): 61.5%
• 35–44 (Millennial mid-career): 47.9%
• 45–54 (Gen X): 40.7%
• 55+ (Boomer): 32%

The pattern repeats by industry. Functions with high applicant volume and structured intake processes, like product management (80.0%), consulting (73.3%), finance (59.4%), customer support (59.1%), sales (57.6%), and software engineering (53.3%), show the heaviest silent-rejection rates.

Healthcare, where licensure and credentialing still pull humans into the loop, sits lowest of the top ten at 44.3%.

Who’s responsible—the human or the machine?

When candidates are asked who they think made the call, the algorithm wins by a wide margin.

Of the 538 candidates who actually received a no-human-feedback rejection, 63.8% blamed an AI. Only 13.9% thought a person was behind it. That’s a 4.6-to-1 split blaming the algorithm.

22.3% say they have no way of knowing how they got rejected. When candidates can’t tell whether a human or a system rejected them, they tend to assume the worst. And once that assumption is in place, every rejection that follows lands on the algorithm by default—whether the algorithm earned it or not.

Only 1 in 10 candidates were told an AI was involved

In theory, the disclosure problem is solved. New York City’s AEDT law, Illinois and Maryland’s video-interview rules, the EU AI Act—all of them require some version of telling the candidate when an algorithm is in the loop.

The candidates we surveyed don’t seem to be hearing it.

Less than 10% of candidates were clearly told an AI was evaluating them. Combined with the 16.2% who don’t know, 84.7% of applicants are operating without a basic answer to the question of who—or what—is reading their resume.

That changes the math of every rejection that follows. A candidate who doesn’t know AI was used can’t ask why it scored them the way it did. They can’t request human review. They can’t even tell whether the disclosure law on the books in their state was followed. The black box stays black.

Why nearly a third of candidates are walking away from AI interviews

Some candidates have stopped waiting to be rejected. Nearly a third (31.4%) told us they’ve walked away from a job altogether rather than sit through a one-way AI video or chatbot screening.

335 of 1,066 chose to give up the role rather than perform for the camera with no one on the other side. And it isn’t the candidates with the most options walking away the fastest. It’s the candidates with the fewest.

The under-$100k middle is absorbing the AI gate

Of the 658 respondents who told us what the abandoned role paid, 79.1% were under-$100k jobs.

Cut by personal income, the gap gets sharper. People earning over $200k abandon AI interviews at just 17.5%—about half the rate of every other band. Candidates with the most leverage rarely face the one-way AI gate. The ones with the least face it the most, and walk anyway.

Younger candidates walk away faster, too: 36% of 18–24s and 35% of 25–34s have given up at least one AI screening, against 21% of those 55 and over.

Put the income and age cuts together, and you get a hiring funnel that’s losing exactly the demographic most employers say they can’t recruit fast enough.

Half of candidates already think AI hiring is biased against them

Whether or not the algorithms actually are biased, the candidates have already made up their minds.

We asked them how much they agreed with this statement: “AI hiring tools are programmed with biases that make it harder for someone of my age, race, gender, or background to get hired.”

47.7% agree to some extent. 25.8% disagree. Almost half of U.S. candidates think the system is rigged against them, and they outnumber the people who disagree by close to two to one.

The bias is loudest among neurodivergent job-seekers

Roughly 23% of our respondents identify as neurodivergent—diagnosed, self-identified, or still figuring it out. Their answers on the bias question stand out.

That gap matters. The behaviors most one-way interview platforms score on—eye contact, vocal pacing, response speed, an even smile—are exactly the behaviors that read differently for many neurodivergent candidates. The system is grading them against a default they don’t fit, and they can tell.

Gen Z and boomers are the most suspicious. Millennials are the least.

The age cut isn’t linear. It bends.

Candidates over 55 and those under 25 see themselves as the most filtered out (the older end reading themselves as too expensive or too senior, the younger end as too inexperienced).

The 35–44 group, who happen to fit the demographic that hiring algorithms have the most training data for, are the least suspicious.

The disclosure paradox: transparency makes things worse

The strangest cut in the data comes when you cross bias agreement with whether the company actually disclosed the AI in the first place.

The more candidates know an AI is there, the less they trust it. Fine-print disclosure produces the highest bias agreement of any group—65%. “I don’t know” produces the lowest, at 35.8%. The closer the disclosure is to a legal hedge, the angrier the candidates get.

Half of candidates are now bringing AI of their own

If you believe the algorithm screening you is unfair, the rational thing to do is bring an algorithm of your own. A lot of respondents have.

It looks like a Gen Z story until you sort by job level

By age, the live-cheating rate descends predictably:

• 18–24: 8.0%
• 25–34: 6.3%
• 35–44: 5.8%
• 45–54: 5.3%
• 55+: 1%

The headline read is the easy one—Gen Z is most willing to game the system. But cross-cut by professional level, the picture flips.

C-suite executives admit to live AI-cheating at 8.6%. Entry-level candidates do it at 1.8%. The top of the org chart is gaming the screening at almost five times the rate of the bottom. Pick your read on why this is happening.

Senior candidates probably encounter more one-way AI interviews to begin with, they almost certainly have more to lose from a bad take, and they may simply be less worried about being caught. Pick more than one. They aren’t mutually exclusive.

Whatever the explanation, the “Gen Z is cheating their way in” framing falls apart on the data. The cheating tracks closer to seniority than to age.

What AIs are actually telling job seekers

At the end of the survey we left an open-ended question: What’s the strangest or most inhuman piece of feedback or instruction you’ve ever gotten from an AI during the hiring process?

Here are themes that came back the most:

None of these are huge in raw count, but the shape of the complaint is consistent. The job seekers who run into an AI job interview tend to describe a system that asks them to perform humanity on a script. Smile here. Look there. Don’t pause. Don’t um. And then it grades them on it.

Below are some pull-quotes from Enhancv’s survey:

Final thoughts: hiring has gone dark

The findings suggest a definitive shift in the candidate–employer contract. Half of U.S. job-seekers are now being rejected by something they can’t see, can’t question, and were rarely told existed. Almost half believe that something is biased against them. Nearly a third refuse to submit to it at all. And among the half who do, a meaningful share are now meeting AI with AI.

For employers, the lesson is that opaque AI hiring is expensive. It hands job seekers the worst-case interpretation of every “no,” pushes lower-paid talent out of the funnel, and breeds the exact mistrust that fuels the next generation of evasion tactics.

Transparency is no longer the polite version of compliance, but the price of running a hiring funnel that candidates will still walk into.