A New York nonprofit hired a grant writer after a strong interview. His answers were polished and confident. Within a month, his manager realized he couldn't do the actual job.
Bloomberg reported the manager now believes he was quietly consulting a chatbot during the virtual interviews the whole time.
Cases like this have pushed employers toward interview monitoring tools that promise to detect AI-assisted answers. The demand has grown quickly, but the results haven't always matched the marketing.
A review of more than 19,000 real interviews found that flagged answers still pass most scoring systems, meaning a strong score alone doesn't rule out AI help.
A high score and a real skill aren't the same thing, and software built to score an answer struggles to tell them apart.
Why a polished answer isn't proof of anything
Part of the problem is how convincing a scripted answer can sound. AI writes in complete sentences and hits every point a question is looking for.
A real answer to something a person hasn't thought about before often sounds rougher, with a pause, a correction, a moment of thinking out loud.
Interviewers are trained to reward confidence and polish, and that's usually the right instinct.
A scripted answer is built to take advantage of exactly that instinct.
The grant writer from the Bloomberg story wasn't caught because his answers were bad. He was caught because his work, once he had the job, didn't match how confident he'd sounded getting it.
What experienced recruiters were already doing
Recruiters have always looked beyond the first impression.
A resume has never told the whole story either, since most people write their CV to make their best case, not to give a neutral account of their work.
Job-hopping used to be treated as a red flag on its own. These days, the reason behind a move carries more weight than how often someone changed roles. Someone leaving to grow their career sends a very different signal than someone leaving because they're disengaged, and the only way to tell those apart is to ask.
The same logic applies to how a candidate behaves during the process itself, not just in the interview. How quickly someone responds, whether they're easy to reach after a call, whether they follow up without being chased, says something real about how they'll behave once they're doing the job.
Good recruiters were already working this way before AI entered the picture. Taking a single data point, on paper or on a call, as the whole picture has never been reliable.
Where AI fits into the process, and where it doesn't
AI has a real place in hiring. It can speed up parts of the process that don't need judgment.
The final call is a different matter, and that's exactly where paying closer attention has become more important, not less.
It's not unusual to hear about tools built to feed candidates answers during a live interview. That's part of why LevelUp treats small details as seriously as it does.
LevelUp core value:
We Sweat the Small Stuff
[quoted: LevelUp core value: We Sweat the Small Stuff]
Something that looks minor in the moment is exactly the kind of thing that tends to turn into a real problem later if it goes unchecked.
What this looks like in a real interview
A prepared answer handles a broad, expected question well. It has more trouble with a specific follow-up nobody could have predicted in advance, something like "walk me through what you'd do first" or "what would change if the client pushed back on that."
Tone gives it away too. A live conversation drifts, circles back, and jumps between topics without warning. A candidate relying on a script often needs an extra beat to catch up when the conversation moves somewhere unexpected. Because whatever's generating their answers is still working through the last question when a new one comes in.
Catching this doesn't need special software. It needs an interviewer willing to leave the script and follow where the conversation goes.
Human judgment isn't perfect either
Recruitment has never been perfect. Recruiters get it wrong too, and no interview process catches everything.
Some candidates give outstanding interviews and leave within a few months. Others were quiet and understated in an interview but they turn into some of the most reliable long-term hires.
That’s why no experienced recruiter relies on one signal. Experience, motivation, communication, references, and the interview all contribute to the decision.
One good interview rarely shows whether someone's motivated or genuinely a fit. Response patterns and follow-up conversations end up carrying just as much weight.
Questions employers are asking
Does AI make interviews less trustworthy?
No. Most candidates using AI are preparing, not lying, tightening a resume or rehearsing answers, which is a different thing from reading live responses off a screen mid-call. Suspicion isn't the goal. Asking questions specific enough that the difference surfaces on its own is.
Should candidates avoid AI when preparing for interviews?
Using AI to prepare is fine, and most people do it now. The real problem is using it live, during the call itself. This is more noticeable once the conversation goes off script.
Doesn't detection software solve this more reliably than a person can?
Not on its own. Software can spot a pattern in someone's answers, but it can't ask a follow-up question or judge the response that comes back. Modern recruitment processes use both: software to flag something worth checking and a person to actually check it.
What an interview is for
AI changed how candidates prepare for an interview. It didn't change what an employer is trying to find out.
A good interviewer has always noticed when an answer doesn't quite fit, and has always been willing to follow a conversation somewhere the plan didn't cover.
A confident answer used to be a fairly reliable signal on its own. These days it takes more digging to know what's behind it, and noticing that gap matters more than it used to.
A routine question asked in a predictable order rarely reveals much. Asking why, or pushing into what happens next, usually does.
Good interviewers rely on questions like these to find out whether someone can do the job, not to catch anyone in a lie. Nothing about that approach is new. It's carrying more of the load these days.
For offshore hiring, interviews often carry more weight, since managers don't have the informal day-to-day interactions that sometimes help validate an early impression in a local office.
If you're building an offshore team, the same principle applies throughout the hiring process.
