AI Interview Candidate Experience: What Job Seekers Say and How to Fix It

AI interview candidate experience is now a hiring metric, not a side issue. Greenhouse's 2026 survey of 2,950 job seekers found that 63% have already faced an AI interview, and 38% say they have withdrawn from a hiring process because of one. If you use an AI interviewer to screen candidates, the way you introduce it, explain it and follow up decides whether strong people finish the process or quietly leave. This guide covers what candidates report, why they drop out and seven changes that keep AI interviews fair and effective.

Table of Contents

• What AI interview candidate experience means
• What the 2026 data shows
• Why candidates drop out
• Seven ways to improve it
• How to measure it
• FAQ
• Conclusion

What AI interview candidate experience means

AI interview candidate experience is how a candidate feels and behaves across an AI-led interview: before it starts, while they answer and after the result. It covers disclosure, clarity of the questions, perceived fairness, the time it takes and whether anyone tells them what happened next.

AI interviewers range from voice-based screeners that ask role-specific questions to pre-recorded video tools that score answers. They solve a real problem for recruiters, because a single person cannot phone-screen hundreds of applicants. If you are comparing formats, our guide to the best AI interview software explains the three categories. This article focuses on the candidate side of the same decision.

What the 2026 data shows

Greenhouse published its 2026 Candidate AI Interview Report in spring 2026, surveying 2,950 job seekers in the US, UK, Ireland, Germany and Australia. Greenhouse sells hiring software, so read it as a vendor survey, but the numbers line up with what recruiters hear in the field. Headline figures, as Greenhouse reports them:

Finding	Share of job seekers
Have experienced an AI interview	63%
Say AI use was not clearly disclosed before their latest AI interview	70%
Have withdrawn from a process because of an AI interview	38%
Would withdraw if an AI interview were required	12%
Want an option to request a human interview	46%
Want confirmation that a person reviews AI output	38%
Got no outcome after an AI interview (never heard back or still waiting)	51%
Have a more positive view of an employer after a good AI interview	38%
Have a more negative view of an employer after a poor one	34%

Two points stand out. First, candidates are not uniformly against AI. Only 19% want less AI involvement than today, while 21% want the same level with more transparency and 22% want more AI with stronger human oversight. Second, perceived bias is similar for AI and for humans, with 36% citing age and 27% citing race or ethnicity in both cases. Candidates are asking for fairness and clarity, not for AI to disappear.

Why candidates drop out

Candidates leave when an AI interview feels hidden, one-sided or silent afterward. Greenhouse lists the top reasons for abandoning a process:

• Pre-recorded, AI-scored video with no human present: 33%.
• No disclosure that AI was being used: 27%.
• AI monitoring during the interview: 26%.
• A required AI-led interview: 26%.

Notice what is missing from that list: the use of AI itself is not the main trigger. The problem is surprise and a lack of human accountability. One in five candidates in the survey said they only found out an AI was involved after the interview began.

Silence compounds the damage. If half of candidates hear nothing after an interview, your AI step becomes another source of ghosting, which our guide on reducing candidate ghosting treats as a process failure rather than a candidate problem.

Seven ways to improve AI interview candidate experience

1. Disclose before the invite, not during the call

Say in the job post or the first email that the first interview is AI-led. Tell candidates how long it takes, what device they need and what happens next. Greenhouse found that 39% want an explanation of what the AI measures, so include that in the same message.

2. Explain the criteria in plain language

Tell candidates the skills you are looking for: for example, problem-solving, relevant experience and communication. Vague scoring feels arbitrary. Structured, job-related criteria also make the process easier to defend if someone questions it.

3. Offer a human alternative

46% of candidates want the option to request a human interview. You do not need to staff a second track for everyone. A simple line such as "Reply if you would prefer to speak with a recruiter" removes much of the pressure and also supports accessibility needs.

4. Keep a person in the decision

An AI interviewer should produce evidence for a recruiter, not a verdict. Say so in your invite. Platforms differ on this point: ConnectDevs' AI interviewer, SAM, returns structured transcripts, summaries and ranked evaluations, and describes humans as making the final decisions. Whatever tool you use, make human review a documented step. We explore the division of labor in AI recruiter vs human recruiter.

5. Make the interview short and job-related

A 10 to 15 minute conversation tied to the role beats a long generic quiz. Ask about the actual work. If you are hiring engineers, a technical screen that adapts to seniority and stack feels fairer than a personality test.

6. Close the loop on a fixed timeline

Promise a response date and keep it. Even an automated "we are not moving forward" message beats silence. With a 51% no-outcome rate across the survey, a reliable reply is a cheap way to stand out.

7. Audit outcomes for bias and fraud

Compare pass rates across candidate groups, and review a sample of rejections by hand. Be aware that AI interviewers do not remove the need for authenticity checks: SAM's own page says it does not claim to detect rehearsed or AI-generated answers, and that standardized scoring does not eliminate all bias. Rules also apply. New York City's Local Law 144 requires bias audits and candidate notice for covered automated employment decision tools, so check with counsel for your locations.

How to measure it

Track a few numbers each month and compare them with your non-AI screening step:

• Completion rate: candidates who finish divided by candidates invited.
• Drop-off point: where in the invite, start or questions candidates leave.
• Time to outcome: days from interview to a response.
• Candidate feedback: a one-question survey after the interview.
• Pass-through by stage: how many AI-screened candidates the hiring manager accepts.

If completion falls well below your phone-screen acceptance rate, the fault is usually the invite or the format, not the candidates. For a wider look at high-volume funnels, see how to screen high-volume applicants.

FAQ

Do candidates dislike AI interviews?

Many dislike poorly handled ones. In Greenhouse's survey, 38% have withdrawn from a process because of an AI interview, yet only 19% want less AI involvement overall. Transparency and human oversight matter more than the technology itself.

Should you tell candidates an AI is conducting the interview?

Yes. Greenhouse found that 70% of candidates said AI use was not clearly disclosed, and 57% think disclosure should be legally required. Some jurisdictions already require notice, so disclosure is both good practice and a compliance step.

Can candidates ask for a human interview instead?

They can if you allow it. 46% of surveyed candidates want that option. Offering it by email reply is simple, and it helps candidates who have accessibility needs or strong objections.

Are AI interviews biased?

They can be. Candidates in the survey perceive similar levels of bias in AI and human interviews, and structured scoring does not remove all bias. Audit outcomes by group and keep a human reviewer in the decision.

How long should an AI interview be?

Keep it as short as it can be while covering the role's core skills. Ten to fifteen minutes is a sensible starting point, then adjust using your completion and drop-off data.

Do AI interviews improve employer brand?

They can in both directions. 38% of candidates say a good AI interview improved their impression of the employer, while 34% say a poor one made it worse.

Conclusion

Candidates are willing to talk to an AI interviewer when it is disclosed, explained and reviewed by a person. They leave when it is hidden, impersonal or followed by silence. Fix the process around the interview first, then choose a tool that supports it.

Want structured first interviews with a human in the final decision? Run an AI interview with ConnectDevs on a live role and see the transcript and summary before you commit.