How to Tell If a Candidate Is Using AI During Interviews

By Cheatproof.ai Team | Published September 4, 2025

In the era of AI like ChatGPT and Cluely, job interviews have become a challenge for employers: candidates often use AI for ready-made answers, undermining the honesty of hiring. In this article, we'll break down the signs of AI usage so HR and interviewers can identify real skills.

General Signs in Speech and Behavior

Pauses and Delays

One of the most noticeable indicators is unnatural pauses and delays in communication. If a candidate responds slowly, with long breaks between the question and answer, it may indicate they're waiting for AI to generate text. A normal person usually starts formulating thoughts right away, even if they're raw, while AI needs time to process the query.

Filler Words and Mumbling

Another red flag is mumbling or filler words at the start of an answer, like "uh," "um," or "well." These sounds are often used to fill a pause while the candidate reads a prompt from the screen. In natural conversation, such pauses happen, but if they appear systematically before every answer, be careful.

Eyes Moving Around

Pay attention to the candidate's eyes: if they move around, not focusing on the camera, it could mean they're looking at another monitor or device with AI. This is especially noticeable in video interviews—the gaze shifts down or sideways, as if reading text from a screen.

Staring Blankly and Sudden Answers

Finally, a classic scenario: the candidate stares blankly for a long time, stays silent or thinks without words, then suddenly delivers a correct, complete, and accurate answer. Without intermediate thinking steps, this looks suspicious, as AI can provide a ready solution in seconds, but a human needs time to ponder.

Signs in Coding

Silent Freezing Followed by Perfect Solutions

When it comes to technical tasks, especially programming, signs of AI use become even more obvious. For example, a candidate might stay silent or "freeze" over a task for a long time, not explaining their thoughts aloud, then deliver a perfect solution that covers all possible edge cases. In real life, developers usually try different approaches, make mistakes, and refine code step by step, but AI generates an optimized version right away.

Overly Perfect Code Structure

Another alarming signal is an overly perfect solution with excellent variable naming, structure, and even comments. The code looks like it's from a professional repository: variables are named logically (e.g., "userInput" instead of "x"), the structure is clear, using best practices. If the candidate doesn't show the creation process (e.g., doesn't comment on why they chose this algorithm), it could be a copy-paste from AI. Real developers often start with a draft where variable names are simple, and the structure is refined later.

Conclusion

Recognizing these signs is key to honest candidate selection. If you notice several of them, don't rush to accuse, but conduct additional checks: ask to explain the code step by step, pose follow-up questions, or change the task on the fly. Tools like CheatProof.ai can automate the process and minimize deception risks. Ultimately, honest hiring benefits everyone: the company gets qualified employees, and candidates get a fair assessment of their skills. Stay vigilant, and AI won't fool your team!