How to spot AI-assisted candidates early.
The 10-Minute Signals in a 60-Minute Interview
How to spot AI-assisted candidates early
Many hiring teams report that sometimes something feels slightly off during remote conversations or technical exercises. Our team has experienced this as well during remote technical interviews.
Instead of discovering it 60 minutes later, you can often spot signals within the first 10 minutes.
Here are 4 practical “authenticity checks” that help.
1. The Small-Talk Script Breaker
Question:
"How is your day going so far?"
or
"What’s the best thing you’ve eaten this week?"
Why it helps:
Candidates prepared with scripted answers often pivot immediately back to their resume or rehearsed talking points (e.g., “I have 8 years of Java experience…”). Natural conversations usually start… naturally.
2. The Subjective Opinion Probe
Question:
"What piece of industry news annoyed you recently?"
or
"What tool in this stack frustrates you the most and why?"
Why it helps:
Real practitioners usually have specific opinions and “battle scars.” Generic or overly polished answers can sometimes signal rehearsed or AI-assisted responses.
3. The Live Walkthrough
Action:
Ask the candidate to explain their thinking step-by-step while solving a problem.
Why it helps:
It’s much harder to rely on external help when someone needs to reason aloud in real time and explain their decisions.
4. The Mid-Exercise Curveball
Action:
During a technical discussion, suddenly add a new constraint.
Example:
"Actually, this system now needs to support 10M users. What changes in your design?"
Why it helps:
This tests real-time adaptability and understanding, not just prepared answers. Candidates relying heavily on external assistance may show noticeable delays or difficulty adjusting their reasoning.
Signals that may be worth a second look
• Repeated requests to repeat simple questions (sometimes used to buy time for external help)
• Long pauses before answering straightforward questions
• Answers that sound polished but lack concrete examples
• Difficulty explaining their own decisions during a live walkthrough
• Eyes repeatedly shifting toward another screen while answering
• Noticeable audio delays or responses that feel slightly out of sync
• Reluctance to adjust camera position or share screen when asked during a task
None of these is proof on its own, but patterns matter.
The goal is to create an interview environment where real experience and thinking become easy to recognise.