How to tell a candidate is using AI in an interview

There is no one sign that proves AI use by a candidate. Interviewers with experience watch for a combination of them, and follow up with questions that outside help cannot easily answer.

What to look for

• Every answer starts with a pause. Before a difficult question a brief pause is expected. When an easy question gets an identical pause and then a smooth reply, something is odd.
• The eyes appear to read. They travel from left to right over a single area of the screen, or go down to one spot ahead of every answer.
• Answers resemble summaries. They are neat, general, frequently structured as lists, and thin on detail from the candidate's experience.
• Follow-ups go badly. After a strong first answer, a plain question such as "why did you choose that?" produces vagueness or a further long pause.
• You hear typing, yet nothing the candidate is sharing is being typed into.
• The answers and the résumé disagree. A subject the résumé hardly touches gets a detailed answer, while the skill listed first gets a brief one.

The tools that produce these signs are explained in how AI interview assistants work.

Questions outside help cannot easily answer

• Pick a decision from the candidate's own work, ask about it, and then ask what they would do differently.
• Once a problem has been answered, alter one of its conditions and ask what happens to the answer.
• Have the candidate explain the answer for a person who is new to the field.
• Keep the question brief and expect a brief reply.

Role-by-role examples are in how to stop AI during an interview with the questions you ask.

When you cannot tell

Record the question, the time and your observation. Afterward, go over the notes with colleagues, and think about a second interview before turning the candidate down. What to do ahead of the interview and once it has ended is set out in how to stop AI use in interviews.