Recruiters Are the New Lie Detectors: AI in Interviews and Resumes

There is a new skill on every recruiter's resume this year, and you will not find it on any certification exam: lie detection. No polygraph, no flashing lights, just a trained ear, a few well-timed follow-up questions, and the growing suspicion that the person on the other side of the screen has a very helpful friend typing just out of frame.

AI has changed how people look for work, and most of that change is good. It helps candidates write cleaner resumes, prep smarter, and stop staring at a blank cover letter. But it has also created a brand-new job for recruiters and hiring managers: figuring out where the candidate ends and the chatbot begins.

AI Is in the Interview Room Now

The numbers tell the story. Resume Genius's 2026 Job Seeker Insights Report, a survey of 1,000 active U.S. job seekers, found that 78% use AI at some point in their search, and 22% admit to using it in real time during interviews. One HR consultant quoted by Newsweek called that 22% likely an underestimate, and from what I see on my desk, I agree.

Resumes are even further along. Novoresume's Resume Report 2026, a survey of 2,000 U.S. adults, found that 42.6% used AI tools the last time they updated their resume. Among people who let AI write most or all of it, 74.8% admitted to including something inaccurate, about three times the rate of people who used no AI at all.

And at the extreme end, Gartner projects that by 2028, one in four candidate profiles worldwide could be fake. In a Gartner survey of 3,000 candidates, 6% admitted to interview fraud, either posing as someone else or having someone pose as them.

What It Actually Sounds Like

Here is a real example from a recent interview. I was speaking with a candidate for an IT infrastructure role and asked a simple scale question: roughly how many mailboxes did the environment she supported include?

Her answer, word for word:

"Exact mailbox count isn't specified in my resume, so I clarified the scale of the environment rather than give an inaccurate number."

Read that again. Nobody talks like that. A person who managed an email environment would say something like "I think it was around 4,000, give or take." That answer sounds like a chatbot that was handed my question and her resume, noticed the number was missing, and politely explained its reasoning. She read it aloud, rationale and all.

It was not the only moment. There were pauses before simple questions, answers that arrived perfectly structured in three neat parts, and very little that sounded like lived experience. None of it proves anything on its own. Together, it painted a clear picture.

Tailored to the Point of Fiction

The resume side is trickier, because using AI to tailor a resume is smart, and I encourage it. Novoresume found that people who keep four or more versions of their resume average about twice as many interview invitations as people who send the same one everywhere.

The problem is when tailoring turns into mirroring. When a resume echoes the job description keyword for keyword, every requirement magically checked off, the document stops telling you about the person. It starts telling you about the job posting. Then the interview becomes the first real test of whether any of it is true.

That is why so many companies are bringing back a human, face-to-face step. Google CEO Sundar Pichai said in 2025 that Google would introduce at least one round of in-person interviews for certain roles. Candidates seem fine with it, too: a separate Gartner survey found 62% of candidates were more likely to apply for a job that required in-person interviews.

Practical Tips: How to Spot AI Answers (and Stay Fair)

The goal is not to catch people. It is to get to the truth about what someone can actually do. A few things that work:

• Ask for the story behind the number. "How many servers?" is easy to fake. "Tell me about the worst outage in that environment and what you did first" is not.

• Follow up twice. AI is great at the first answer. It struggles with "What happened next?" and "Why did you choose that over the alternative?"

• Listen for AI tells. Restating the question, explaining their reasoning out loud, referring to "my resume" like a separate document, and formal hedging all stand out in a casual conversation.

• Watch the rhythm. A long pause on a simple question, eyes drifting to another screen, then a polished three-part answer is a pattern worth noticing.

• Use the resume as a map, not a script. Pick one bullet and go deep. Real experience has texture: names of tools, mistakes, lessons, small details nobody would invent.

• Add a live step for finalists. Camera-on technical walkthroughs, shared-screen problem solving, or one in-person meeting for local candidates.

• Do not penalize polish. A well-written resume is not a red flag. A resume the candidate cannot explain is.

Why AI Will Not Replace This Part of the Job

Here is the irony: AI created this problem, and AI cannot fully solve it. Software can flag patterns, but it cannot sit across from someone, sense that an answer does not match the person, and gently push until the real story comes out. That takes judgment, rapport, and a lot of conversations.

That is the part of recruiting I love most, and it is exactly why a human-led process matters more now, not less. At High Sierra Talent, every candidate I present has been through a real conversation with me, follow-up questions and all, so you meet people whose experience holds up.

If your team is seeing more polished resumes and fewer people who can back them up, let's talk. Reach me at heather@highsierratalent.com and I will help you find IT talent that is the real deal.