You just got an invite to an AI-led interview, maybe a one-way video with a countdown timer, and your first instinct is to close the tab. A Use.AI survey of 6,731 job seekers, reported by TechRadar, found that 29% of employed candidates have pulled out of a hiring process over an AI interview, compared with 14% of unemployed candidates. Before you join the people walking out, check whether you actually have the leverage they have.
Should I do an AI interview?
Sit the AI interview if you don't hold an offer or a late-stage process somewhere else, unless one of a short list of hard red flags shows up. Most of the people walking out can afford to, and every one who leaves makes the stage less crowded for you.
Two definitions first. An AI-led interview is a screening conversation where software asks the questions, and sometimes scores your answers too. A one-way video interview is a recorded session where you answer prompts on camera with nobody watching live, and the recording is reviewed afterward by a person, an algorithm, or both.
This format is now standard. Greenhouse's 2026 Candidate AI Interview Report surveyed 2,950 active job seekers across the US, UK, Ireland, Germany and Australia. 63% said they had been interviewed by an AI, up 13 percentage points in six months. A January LinkedIn survey of more than 6,500 HR professionals found two-thirds of recruiters plan to use more AI for pre-screening interviews this year. If you refuse the format across the board, you're refusing a large and growing share of the market.
Why employed candidates walk out at twice the rate
Employed candidates withdraw more often because they still have a paycheck to fall back on. The gap between 29% and 14% (2.07x) tells you how much leverage each group has. It doesn't tell you how much either group dislikes the format.
The same survey makes this explicit. 58% of people who sat an AI interview despite having reservations said they would refuse one if they had other jobs lined up. So the objection is roughly the same for everyone. What differs is who can afford to act on it. The split between employed and unemployed respondents held in every region covered: the UK, US, Canada, the European Union and Latin America. TechRadar's coverage framed the finding as employed seekers shunning companies that use AI interviews. Read from the candidate side, it means the people with options are leaving the room.
29% vs 14% in Use.AI's survey of 6,731 job seekers across five regions.
One caveat before you build a strategy on this. "Had withdrawn" means the person withdrew at least once, ever. It is not a per-process dropout rate, and the survey doesn't name the employers, the interview products, or when the survey ran. Treat the ratio as a direction, not a precise forecast. In separate reporting, Germany had the highest withdrawal rate, at 42%.
| Metric | Figure | Source |
|---|---|---|
| Employed seekers who withdrew over an AI interview | 29% | Use.AI via TechRadar |
| Unemployed seekers who withdrew | 14% | Use.AI via TechRadar |
| Employed-to-unemployed withdrawal ratio | 2.07x | Derived (29 ÷ 14) |
| HireVue-skilled profiles, US vs UK | 440 vs 55 (8.0x) | Refolk's index; ratio derived |
| Completed an AI interview and never heard back | 51% (vs 28% advanced) | Greenhouse 2026 |
| Founder/CEO titles in top 10 "AI interviewer" headlines (US) | 8 of 10 | Refolk's index |
The competition at this stage gets weaker, not just smaller
If you stay, you face fewer of the employed candidates, and employed passive candidates are often the strongest people in a pipeline. Self-selection takes out the competitors you would least want to face.
Here is some illustrative math. Don't treat it as a measured rate. Start with 100 applicants, half employed and half not. Apply the survey's withdrawal rates: 35.5 employed candidates stay (71% of 50) and 43 unemployed candidates stay (86% of 50). Employed candidates drop from 50% of the pool to about 45% (35.5 ÷ 78.5). That's a small shift, but it moves in your favor.
It's also worth knowing that most candidates don't object to AI as such. In separate survey reporting, only 19% wanted less AI in recruitment. Their requests were narrower: 40% wanted upfront disclosure and 45% wanted the option to ask for a human interview. That matters for the red-flag section below, because it separates format complaints, which you can live with, from process failures, which you can act on.
The leverage rule: when to sit it, when to skip it
Base the decision on what you're holding, not on how the format feels. With no offer, sit it. With a live offer or a final round, it's optional. With a hard red flag, walk away regardless. Leverage, in this context, means the concrete alternatives you have if this process dies: offers, final rounds, a job you're happy staying in.
- No offer, no late-stage process. Sit the interview. You're in the group that withdraws at 14%, so most of your competitors are sitting it too, and the employed ones are thinning out.
- A final round elsewhere, no offer yet. Sit it if the role matches you better than the other process. If it's a weaker match, you can let it go.
- A written offer in hand. Skipping is fine. You're now in the group behind the 58% figure, and spending an hour on a screen for a job you'd turn down isn't worth it.
- Employed and happy enough to stay. Sit it only if the role is a clear step up. Your current job is your leverage, so you don't need to spend it on a screen.
- Any hard red flag (next section). Walk away, whatever your leverage.
Rules 2 and 4 depend on knowing how well you actually match the role, and gut feel is a poor way to judge that. Refolk scores how well your history fits a given posting before you commit an evening to a recorded screen, so "is this one worth it" has a number behind it instead of a mood.
Walking out of an AI interview is a luxury. Decide based on the offers you actually hold.
Which red flags actually justify walking
Two of the four common triggers for walking out justify it: no disclosure of how AI will be used, and prerecorded video scored by AI with no human review. A required AI-led interview is just a format, and AI monitoring calls for a question before you decide anything.
Greenhouse broke down what drives US candidates to drop out:
| Trigger | Share citing it (US) | Verdict |
|---|---|---|
| Prerecorded video scored by AI, no human present | 33% | Ask about human review; walk if the answer is no |
| No disclosure of how AI is used | 27% | Ask once; walk if they won't answer |
| AI monitoring during the interview | 26% | Ask what is monitored, then decide |
| A required AI-led interview | 26% | Format only; sit it |
The disclosure question is also a legal one in some places. Under Illinois's Artificial Intelligence Video Interview Act, an employer may not use AI to evaluate applicants who haven't consented to AI analysis. If an Illinois employer scores your recorded video with AI and never told you, that breaks the consent requirement, which is a stronger objection than personal preference. Note the employer and the date if that happens. The law has limits, though: it has no bias-audit mandate and no stated penalty.
Scoring with no human review has a public warning attached. In March 2025, the ACLU of Colorado filed a complaint with the Colorado Civil Rights Division and the EEOC against Intuit and HireVue. It alleged that HireVue's AI video interview platform discriminated against a deaf and Indigenous employee. If you have a disability, an accent, or a speech pattern that a scoring model might misread, asking for an accommodation or a human interviewer is a reasonable request. 45% of candidates surveyed want exactly that option.
Who is actually on the other side of the camera
AI interviews are run by a mix of established platforms like HireVue inside large employers and a crowd of founder-led startups whose quality varies widely. That's why "is a human involved?" is a fair question to ask, not a hostile one.
In Refolk's index of professional profiles, 440 US profiles list HireVue as a skill, against 55 in the UK, a ratio of 8.0x. In a sample of 25 of those US profiles, EY and Deloitte were the most common current employers (3 each). The titles were mostly recruiters, talent acquisition directors and HR technology analysts. The established tools sit inside big firms, run by recruiting and HR technology staff.
The newer end of the market looks different. 312 US profiles in Refolk's index have "AI interviewer" in their headline, and 8 of the top 10 titles are Founder, Co-Founder or CEO variants. The employers include Xobin, InterspectAI, GenieHire.ai, Classet and candidate.fyi. The sample breakdowns are directional (samples of 25), but the totals are solid, and together they suggest a fragmented field of founder-led vendors.
Job seeker Art Hebbeler put it to CNBC this way: "We have AI tools talking to AI tools, not people talking to people." That's fair, and it's why the useful move is to find out where a person comes back into the process, not to boycott the tool.
The real risk is silence, not the format
The bigger danger after an AI interview is never hearing anything back, not doing badly in the interview itself. Among candidates who completed one, 28% moved forward, 13% were formally rejected and 51% got no response at all.
Only 28% advanced and 13% got a formal rejection, per Greenhouse's 2026 report.
Greenhouse CEO Daniel Chait has said that "most AI in hiring today is making a bad system worse: more applications, less signal, and less transparency." Silence is what that looks like from your side. Twenty-eight percent advancing is a real path, so sit the interview, but don't stop applying while you wait. The costly mistake is pausing your search for a week waiting on one screen.
Running a parallel pipeline used to mean rewriting your resume for every posting, which is why people stall after an interview. Refolk does that work for you: paste the posting and you get your own resume back, rewritten for that role, plus a cover letter draft, so applying to five more roles while one sits in limbo takes an hour instead of a weekend.
It's worse when you're employed. Every recorded screen is an hour taken from your actual job, half of them end in silence, and you're also trying to keep the search away from your current manager. This is where the 29% come from: people with a job decide the expected return isn't worth the exposure or the hour, and they're often right.
FAQ
Can I decline an AI interview without hurting my chances?
You can decline, but declining generally takes you out of that process. If you hold no offer or late-stage process, declining usually costs more than it saves, because the format is now mainstream (63% of surveyed seekers have done one). A better move than declining outright is to ask for a human interview or for disclosure of how the AI is used. Candidates want those two things far more often than less AI overall (45% and 40%, against 19%).
Is a one-way video interview worth it if I'm currently employed?
It's worth it when the role is a clear step up and you understand how the recording will be reviewed. If you already have a written offer, or the role is a sideways move, skipping is reasonable. That's what the 58% of reluctant sitters say they would do with other options in hand. Your current job is your leverage, so spend the hour only on roles that justify it.
What should I do if I never hear back after an AI interview?
Assume nothing and keep applying. 51% of candidates who completed an AI interview never heard back, against 28% who advanced, so silence is the most common outcome and not a verdict on you. Send one follow-up to a human contact if you have one, then put your energy into the next applications rather than waiting on this one.