Ever since artificial intelligence (AI) chatbots became mainstream, conducting tests and interviews has become a nightmare for recruiters due to rampant fraud. California-based technology leader and research scientist Lily Zhang recently shared a first-hand account of this growing problem on social media. While interviewing for a machine learning (ML) system design, Zhang caught a candidate using deceptive tactics of fabricating answers without relying on his or her intelligence.
According to Chan, she used AI to catch the candidate’s misconduct fairly early on in the interaction, but decided to collaborate to see how the situation would develop.
“Today, I discovered a candidate cheating during an ML system design interview. He was talking without thinking. And I knew it in the first 3 minutes, but I let him speak,” Zhang said in a post on X (formerly Twitter).
After the first part of the interview, Zhang ordered the candidate to share his screen for the test, but the candidate quickly “froze up” and eventually revealed himself.
“We went through the entire design end-to-end. Then I invited him to share his screen and draw the architecture. He froze. Then I saw the chatbot window flicker. I had no choice but to end the interview,” Chan said.
“What are cheaters thinking? Why not buy more reliable cheating software? Isn’t it obvious?”
Check out the viral post here:
Today, I discovered a candidate cheating during an ML system design interview. He was speaking without thinking. You can tell that in the first three minutes, but I let him speak.
We considered the entire design end-to-end. Then I invited him to share his screen and draw.
— Lily Zhang (@lily_gpupoor) July 24, 2026
Also read | Ukrainian woman reveals the difference between Indian and European housewarming parties: “A completely different world”
social media reaction
As the post gained traction, social media users wondered what such a candidate intended to accomplish if caught while on probation.
One user said: “What I’ll never understand is, even if I manage to get in, how are they going to get past the probationary period?” Another added: “I would have stopped at the first red flag. Leaving it like this just wastes everyone’s time.”
A third commented: “The flickering chatbot window is real, not freezing. It means he gave up trying to hide it minutes before you actually called.”
A fourth said: “At a previous organization, candidates would clearly type out questions during interviews, pause to answer, and read out the screen. It was so blatant. I’m still confused as to how they thought they could get away with this.”
