Candidate experience AI hiring: candidates accept an AI evaluator, not a silent decider

The candidate experience AI hiring question has a clearer answer than the debate suggests. Candidates don't reject AI in hiring, they reject AI that decides about them without telling them. The evidence points one way: people are fairly open to AI that evaluates them consistently, and they push back hard on AI that makes the final call.

That distinction changes what you should build. "Be transparent and keep a human in the loop" is advice everyone already gives, and it doesn't tell you what to do on Monday. The useful version is stage specific: where AI helps the candidate, where it costs you their trust, and what they should get back when it's over.

Key Takeaways

Candidates object to AI as a silent decider, not to AI as an evaluator. Design every stage of the process around that line.
Pew Research Center's December 2022 survey of 11,004 US adults found 71% opposed AI making final hiring decisions and 7% favored it, yet 47% thought AI would be better than humans at treating all applicants the same way (15% said worse).
Applicants who had actually been through AI recruitment rated it useful and easy to use, with faster responses as the biggest benefit (Horodyski, 2023).
The cheapest trust win is disclosure at the invite: say AI is involved, what it scores, and who makes the decision.
Fairness to honest candidates now includes catching the ones who cheat, and that monitoring should be disclosed too.

What candidates actually object to

Look at the reasons people give and a pattern shows up. In Pew's survey, the people who said they wouldn't apply to an employer using AI talked about it missing the human factor, missing nonverbal cues, carrying its designers' biases, and passing over people with nontraditional work histories. Those are worries about judgment, about who decides and whether anyone checks.

That's a narrower complaint than "candidates hate AI," and a far more useful one. You can't make an AI feel human, and you shouldn't try. What you can do is keep a person accountable for the decision, tell candidates AI is involved, and show them what was scored.

To be clear about the evidence, Pew didn't test disclosure, so that part is our position rather than a survey finding. But the reasons people give suggest candidate perception of AI hiring turns on judgment and accountability more than on the technology itself. Leave those questions unanswered and even a fair, well-built process reads as a black box, because from the outside it is one.

What the survey data says, and what it does not

The most cited source here is Pew Research Center's survey of 11,004 US adults, fielded December 12 to 18, 2022 and published in April 2023. The headline is harsh. Americans opposed AI making final hiring decisions by 71% to 7%, and a 41% plurality opposed AI reviewing job applications.

On the personal question, 66% said they would not want to apply to an employer that used AI to help make hiring decisions, against 32% who would. Women were more wary (70% would not apply) than men (61%). If you stopped reading there, you'd conclude that AI and a good candidate experience can't coexist.

Keep reading. In the same survey, 47% said AI would be better than humans at treating all applicants the same way, while 15% said it would be worse. Where people trusted AI less was judgment: they expected it to do worse at seeing potential in applicants who don't perfectly fit the description.

Two caveats matter. This is December 2022 data, now nearly four years old, so treat it as a baseline rather than current sentiment. And it asked the general public about a hypothetical, which is a different thing from asking people who have just been through an AI hiring process.

The employer side had its own warning. SHRM's February 2022 survey of 1,688 HR professionals found 64% said their AI or automation tools automatically filtered out unqualified applicants. Only 2 in 5 employers buying from vendors called the vendor very transparent about bias protection, and 19% of organizations using these tools said they had accidentally overlooked or excluded qualified people.

Candidate experience AI hiring: evaluator versus decider

Now set Pew beside people who have actually been through it. Horodyski (2023), in Computers in Human Behavior Reports, studied applicants' own experiences of AI-enabled recruitment and found they perceived it positively, as useful and easy to use. Reduced response time was the biggest benefit. The biggest drawbacks were AI missing the nuance of human judgment, low accuracy and reliability, and immature technology.

Treat that as directional. The abstract doesn't report the sample, and it's 2023 research on AI recruitment broadly, not on AI interviews specifically. Still, it points the same way as Pew: people like AI for speed and consistency and distrust it on judgment calls.

Put the two together and the line isn't AI versus human. It's AI as the evaluator, scoring every candidate against the same rubric and showing its reasoning, versus AI as the silent decider, rejecting people nobody on the team ever looked at. Pew's respondents opposed AI making the final hiring decision by roughly ten to one (71% to 7%), while speed and consistency are where the evidence shows a benefit.

This is also why the recruiter's job doesn't disappear. Someone has to own the decision and be able to explain it, which is the argument we made in will AI replace recruiters. An AI that evaluates gives that person better evidence. An AI that decides removes them, and that's exactly the version the public says it doesn't want.

Stage by stage: where AI helps the candidate and where it costs trust

AI recruiting candidate experience isn't one thing. It's a handful of moments, and AI helps at some of them and hurts at others. Here's how the evaluator versus decider line plays out at each stage.

Resume screening

This is where the silent decider usually lives. Automated filters reject people before a human sees them, and the candidate gets nothing back beyond a form email. If you use AI here, use it to rank and surface rather than auto-reject, and keep a human reviewing the cut line. Blind hiring practices help too, because they limit what the screen can react to in the first place.

Scheduling and response time

Speed was the top benefit in Horodyski's study, and this is where candidates feel it. Nobody objects to a faster invite or an interview they can book the same week. A slow loop costs you candidates whether or not AI is involved, which is why reducing time to hire is usually the quickest trust win available.

The first-round interview

This is the stage people picture when they worry about AI interview candidate experience. A structured AI interview asks every candidate the same core questions against the same rubric, which is exactly the consistency Pew's respondents expected AI to be good at. The trade-offs against a human screen are real, and we've laid them out in AI interview vs human interview.

Two things make it feel less like a black box. The first is a real back-and-forth voice conversation rather than a recorded monologue, so the candidate gets follow-up questions instead of talking into a void. The second is a human who can step in: on Expert Hire, a recruiter can join a live AI interview as a silent observer and take over mid-interview (availability depends on plan).

Scoring and the shortlist

The AI can score and rank the pool, but a person should read the evidence before anyone is cut. That split is the whole point. A rubric score with the transcript excerpt and written reasoning behind it is evidence a recruiter can check, question, and overrule.

The final decision and the rejection

Keep the final decision human, and make that visible. On Expert Hire, final-round interviews can switch to a mode where the AI steps back and a human interviewer takes the seat. When someone is rejected, tell them a person made the call and, where you can, what the process measured.

Telling candidates what was scored

The cheapest candidate experience AI fix is disclosure at the invite. Before the interview, candidates should know three things: that AI will conduct or score part of it, what the rubric covers (problem solving, code quality, and communication, for example), and who makes the decision afterwards.

In some places this isn't optional. For Illinois-based roles, the Artificial Intelligence Video Interview Act (820 ILCS 42) requires employers that use AI to analyze applicant-recorded video interviews to notify applicants, explain how the AI works and what general types of characteristics it evaluates, and get consent beforehand. Expert Hire has a built-in consent flow for it, and the same notice is good practice everywhere.

Then give something back. Where the employer enables it, candidates on Expert Hire can see the criteria and the rubric they were measured against. For the hiring team, every score carries the transcript excerpt and the AI's written reasoning, which is what makes it explainable rather than just a number. We walk through that mechanism in how AI interviews are scored.

Consistency is worth explaining, not just claiming. Schmidt and Hunter's 1998 meta-analysis put structured interview validity at .51. Sackett, Zhang, Berry and Lievens (2022, Journal of Applied Psychology) revised it to roughly .42 after showing earlier estimates overcorrected for range restriction, and structured interviews still came out as the top-ranked selection procedure, ahead of unstructured ones. Our methodology page documents the IO psychology research behind our rubric design and scoring.

Candidates can also practise first. Expert Hire's mock interview mode uses the same format, so someone nervous about talking to an AI can try it before it counts. A surprise format is a poor test of anyone's skill.

Fairness to honest candidates when others cheat

Here's the part of candidate experience most guides skip. If some candidates are reading AI-generated answers off a second screen, the honest ones are being compared against them. A process that can't tell the difference is unfair to the people doing it straight, however friendly it feels.

So proctoring is part of fairness, not just security. Expert Hire's secure desktop proctoring (included on the Growth plan and above) covers tab-switch awareness, screen monitoring, and voice-consistency checks, and a live conversation with live coding makes a scripted answer harder to sustain. Whether candidates may use AI at all is a policy choice, and we've covered the options in should you allow AI in job interviews.

Apply the same disclosure rule to monitoring as to scoring. Tell candidates what's monitored and why before they start. Surprise surveillance is the fastest way to turn a fair process into a hostile one.

Frequently asked questions

Do candidates like AI interviews?

The evidence is mixed and depends on who you ask. Pew's December 2022 survey found most Americans opposed AI making final hiring decisions, while Horodyski (2023) found applicants who had been through AI recruitment rated it useful and easy to use. Neither study looked at AI interviews specifically, but together they suggest candidates accept AI as a consistent evaluator far more readily than as the decider.

Should you tell candidates you're using AI in hiring?

Yes, at the invite, not buried in a privacy policy. Say that AI is involved, what it scores, and who makes the final decision. For Illinois-based roles, notice and consent are legally required when AI analyzes applicant-recorded video interviews.

Does AI make hiring fairer for candidates?

It can make evaluation more consistent, because every candidate gets the same questions and rubric, and 47% of Pew's respondents expected AI to be better than humans at treating all applicants the same way. Consistency isn't the same as fairness, though. The rubric still has to measure the job, and a human should review the outcomes.

Can candidates see their AI interview results?

It depends on the employer. On Expert Hire, candidates can see the criteria and rubric where the employer enables it, and every score carries the transcript excerpt and written reasoning for the hiring team. Candidates who want feedback on their own performance can use the mock interview product, which returns a report card with reasoning per criterion.

Should AI make the final hiring decision?

No. The public opposed it 71% to 7% in Pew's survey, and nothing about a faster process justifies that trust cost. Let AI evaluate and rank, and keep a named person accountable for the hire.

Candidates accept an evaluator they can see

Good candidate experience in AI hiring doesn't mean hiding the AI or apologizing for it. It means using AI where the evidence says it helps (speed and consistent scoring), keeping it out of the silent final decision, and telling people what was scored. Most of that costs you a paragraph in the invite and a clear rule about who decides.

If you want to see what the evaluator side looks like in practice, look at how Expert Hire's AI interview platform produces a scored report card with the transcript behind every score. That's the record a recruiter should be able to explain to any candidate who asks.

By TK, Growth at Expert Hire. Last updated September 29, 2026. Reviewed by Anand Suresh, CPO at Expert Hire.