AI Academy/Module 7
How Should You Handle Candidates Who Use AI in Interviews and Assessments?
Candidates now use AI assistants in live interviews and tests. Here is how to set clear rules, design interviews AI cannot fake, read the warning signs fairly and handle accommodations.
- 7module
- 4 of 5lesson
- 10 minread
The short answer
Set clear rules, tell candidates before the interview, and design your interviews so that AI help does not change the result. Ask follow up questions that need real experience, allow AI where the job itself uses AI, and use in person or supervised steps for the skills that matter most.
Be careful about accusing anyone. The signs of AI help look a lot like nerves, a disability or a bad internet connection, so ask questions rather than jump to conclusions.
How are candidates using AI in interviews?
Tools now listen to a video interview and show suggested answers on screen in real time. Others solve coding tests or answer assessment questions. Some candidates simply keep a chatbot open in another window.
Employers have noticed. Amazon told candidates not to use generative AI tools during interviews unless explicitly permitted, warning that breaking the rule may lead to disqualification. Google banned AI tools in virtual interviews, and Google, Cisco and McKinsey brought back in person interviews for some roles. The same report cites a Gartner survey in which 72.4% of recruiting leaders said they now hold interviews in person to combat fraud.
Is it always cheating?
No. It depends on your rules and on the job. A Cisco talent acquisition leader put it simply: if a candidate is not explicitly invited to use AI in an assessment, it should be treated as off limits. That means the fairest thing you can do is say clearly where AI is allowed.
| Stage | A common rule | Why |
|---|---|---|
| Preparing for the interview | AI allowed | Research and practice make for better interviews |
| Live interview | No AI unless invited | You want to hear how they think |
| Take home task | Say which way, every time | Some jobs expect AI use, some do not |
| Skills test for a job that uses AI daily | AI allowed, and assessed | How well they use AI is part of the skill |
Anthropic publishes a candidate AI policy along these lines and invites other employers to adapt it.
How do you design interviews that AI cannot fake?
| Technique | How it works |
|---|---|
| Ask why, then ask again | AI can give a good first answer. Real experience shows in the second and third follow up |
| Ask about their own work | Questions about a project on their resume need details only they know |
| Change the problem midway | Add a new constraint and ask how their answer changes |
| Talk through a work sample | Let them use AI on a task, then ask them to explain and defend every choice |
| Use in person or supervised steps for key skills | Keep this for the final stage so remote candidates are not put off early |
One outsourcing firm told Computerworld that candidates started acing a quiz on an unusual coding library but could not explain their answers when asked. That is the pattern to design for. A good follow up question does more than any detection tool.
What are the warning signs, and how reliable are they?
Amazon’s guidance to its interviewers listed signs such as typing while being asked a question, appearing to read answers, and eyes moving across the screen. These are worth noticing, but they are not proof.
| Sign | Could also be |
|---|---|
| Pause before every answer | A slow connection, thinking, or answering in a second language |
| Eyes moving to one side | Notes, a second monitor, or a habit |
| Very polished, generic answers | Heavy preparation |
| Typing during questions | Taking notes, or a disability that needs notes or captions |
If you are unsure, ask a follow up question about their own experience, or politely ask them to close other windows. Do not accuse. Some candidates use assistive technology as a disability accommodation, and US law requires employers to consider reasonable accommodations in hiring. Tell candidates how to ask for one before the interview.
Is this the same as interview fraud?
No. A real candidate getting AI help is a question of fairness. Someone pretending to be another person, or using a deepfake, is fraud and a security risk. Module 3, Lesson 5 covers how to spot and stop it.
Quick answers
Should I record interviews to catch AI use?
Only with the candidate’s consent and within recording laws where you and they are. Recording does not stop AI use on its own.
Can I ask candidates to share their screen?
For a technical task, yes, if you tell them in advance and treat everyone the same way.
What if I find out after the offer?
Talk to HR and legal before acting. What you can do depends on your stated rules and local law.
Do AI interviewers have the same problem?
Yes, often worse, because nobody is there to ask a follow up. See AI interviewers in Module 3.
Try it on a live req this week
Tell candidates on a live req, in writing, what AI use is and is not allowed in interviews, and add one follow up question per interview that goes deeper on their own example.
Explore in the directory
Module 7: When candidates use AI too
- How are candidates using AI to apply for jobs?
- What do you do when auto apply bots flood your job posts?
- Can you tell if a resume was written by AI, and does it matter?
- How should you handle candidates who use AI in interviews and assessments?
- Will AI replace recruiters, and which skills keep you valuable?