Nobody Hires the Perfect Answer

"That's a good question."

If you hear those four words, then a short pause, then an answer that sounds like a press release, you may be talking to a chatbot with a human face in front of it. At least that is what recruiters in the UK are now saying out loud.

The numbers from this week

This week, the Institute of Student Employers in the UK published its yearly survey of graduate recruiters. 92% of employers said they suspected or confirmed misconduct from candidates. Two years ago it was 78%.

The most common problem was undeclared AI use during interviews, reported by 66% of employers. Using AI in online assessments came next at 47%. Using it for CVs and cover letters? Only 10% saw that as an issue, and most employers now openly accept it.

Chris Morrow, who runs a recruitment agency in Manchester, told The Guardian how he spots it. The candidate stalls while the tool generates something, fills the gap with stock phrases, and sounds strangely formal. "It's like they are reading off an Autocue," he said. His agency simply ends the interview.

The ISE's chief executive, Stephen Isherwood, put the real risk in one line: if you misrepresent yourself, you could end up doing the wrong job.

What the cheaters misunderstand

I get the pressure. I have been the nervous candidate many times, in Lebanon and later in Germany. When you are interviewing in a second language and someone asks something hard, a perfect answer whispering in your ear sounds like a gift.

But here is the thing. An interview is not an exam with a right answer at the back of the book. Nobody remembers the perfect answer. What people remember is how you think when you don't have one.

An answer like "I'm not sure, but here is how I would find out" tells an interviewer more than any polished paragraph. It shows how you reason. It shows if you are honest about the edges of what you know. That is exactly what your future team will need on a bad Tuesday when production is down and there is no script.

An interview is not a test of the answer. It's a preview of what it's like to work with you when you don't have one.

A tool can give you the words. It can't give you the moment where you hesitate, think, and land somewhere real. And that moment is the product.

Where this is going

Employers are already reacting. According to the survey, 77% are reviewing how they recruit. More personal questions, time limits, randomised assessments. My guess is we will also see more in-person rounds again, more "explain this code you wrote" conversations, and fewer generic questions that a model can answer in two seconds.

The report is about the UK, but I don't think Germany is far behind. Video interviews are normal here too, and so are the tools.

What we can actually do

If you are interviewing:

1. Practise saying "I don't know yet" followed by your plan. Out loud, with a friend. It feels awkward the first time and very natural by the fifth.
2. Use AI to prepare, not to perform. Let it quiz you, challenge your CV, play the tough interviewer. Then close it.
3. Bring real stories. A model can't invent the bug you fixed at 2 a.m. or the conversation with a difficult colleague.

If you are hiring:

1. Ask follow-ups. The first answer can be borrowed. The third "why" rarely can.
2. Ask about their own work: a project, a decision, a mistake. Specific beats generic.
3. Write down your AI policy and tell candidates before the interview. Most people want to play fair when they know the rules.

Give me "let me think" and a few seconds of real silence over a flawless answer that arrives too late and sounds too clean. That pause is worth more than any perfect answer.

Sources: The Guardian via AOL: Use of AI in job interviews on the rise, HRreview: AI misuse in graduate recruitment reaches record high, Resultsense: 92% of UK graduate employers report misconduct