Using AI in Interviews Ethically: Where the Lines Are

Three bright lines, a wide gray zone, and the disclosure norms that resolve most of it. For candidates and for the companies who hire them.

30-SECOND TAKEAWAY

The bright line: real-time, undisclosed AI assistance during a live interview is fraud. No matter how the marketing positions it.
The other bright line: preparation with AI is fine and expected. You don\'t owe the interviewer a list of your prep tools any more than you owe them your flashcard count.
The gray zone is take-homes and screen-shared coding. The default move when the company is silent: ask. Most ambiguity comes from companies that didn\'t specify and didn\'t mean to leave it open.

Three bright lines

Preparation with AI — fine

Drilling questions with ChatGPT, using AI mock interviews, having Claude critique your draft answers, asking the model to explain a concept: all fine. The interview measures what you can do; what tools you used to prepare to do it are not the interviewer\'s business. You do not list your prep tools on your resume and you do not need to disclose them in the interview. Reasoning: the meta-skill being measured is "did you prepare," not "how did you prepare."

Real-time covert AI in live interview — fraud

Using Final Round AI, LockedIn AI, Cluely, or any equivalent tool to generate answers during the live interview, without disclosure: fraud, in any reasonable reading. The interview is a representation about your capabilities; allowing an AI to manufacture answers in real time is a misrepresentation. Reasoning: the company is paying you for the skills they observed in the interview. If those skills are not yours, the offer is procured by deception, and offer rescissions follow if the truth ever surfaces.

AI-generated cover letter submitted as your own — deceptive

Asking ChatGPT to write a cover letter and submitting it verbatim: deceptive but increasingly normalised. Asking ChatGPT for a draft and then editing it heavily into your own voice with your own specifics: acceptable. Asking ChatGPT for "a cover letter" and pasting the output unchanged: lazy, detectable, and signals low effort to anyone who has read more than ten of these. Reasoning: the cover letter is your writing as a sample; using AI as a research assistant is fine, using it as a ghostwriter without acknowledgement is closer to plagiarism than to preparation.

The gray zone, decoded

Async take-home tests

If the company stated a no-AI policy in writing and you used AI, that is cheating. If they were silent and the role assumes AI fluency on the job (most engineering roles in 2026), using AI in the take-home is reasonable — declare it in the submission and briefly explain how. If the company explicitly allowed AI assistance, no issue. Default move when policy is unstated: ask the recruiter. The 30-second email saves a much bigger problem later.

Screen-shared coding with browser tabs open

Most companies expect the candidate not to be reading from ChatGPT or Stack Overflow during a screen-shared coding interview. If the interviewer can see your screen and you keep a tab visible with the answer, you have signalled either honest hands-on practice or unconcealed cheating, depending on the company\'s norms. Ask up front: "Is it okay if I keep documentation open in another tab?" The yes-or-no settles the question.

Asking ChatGPT to verify answers after the interview

Completely fine. The interview is over; you are reflecting on how you did. This is the standard learning loop and using AI to accelerate it does not retroactively affect the interview itself. Useful prompt: "Critique the following answer to a senior engineer interview question — what was strong, what was weak, what would a senior interviewer push back on?"

The decision framework

Stated rules first. Role norms second (a senior IC engineering interview at most companies in 2026 implicitly allows AI in preparation but not during the live session). Ask the recruiter third. If the answer is unclear from any of these and you are uncomfortable: do not use it. The defensive position is durable; the aggressive position requires you to be right about the unstated norm, every time.

For companies: how to specify your AI policy

Take-home AI policy — model language

Pick one of three positions and write it into the take-home brief. Position A: "You may use AI assistance freely on this exercise. We expect you to use it on the job; we want to see how you use it well." Position B: "AI assistance is not permitted on this exercise. We want to assess your unaided problem-solving for this stage." Position C: "AI assistance is permitted; we ask that you note in the README what tools you used and where they helped." All three are defensible. The mistake is leaving it unstated.

Real-time disclosure — model language

Open every live interview with one sentence: "This is a live conversation, and we ask that you do not use real-time AI assistance like answer-generation tools during the call. Standard prep, reference materials, and notes are fine; live AI overlays are not." Spend ten seconds. The clarity prevents most of the ethical confusion and gives you a clean basis for any later concern.

Consequences of undisclosed use

Be explicit in the offer letter. "Any offer extended is contingent on the accuracy of representations made during the interview process, including the candidate\'s direct authorship of answers given." Lawyers can refine the exact language; the principle is that the offer is conditional and undisclosed AI use is grounds for rescission. Most companies bury this in the standard background-check clause; making it explicit is more honest with the candidate and stronger if you ever need to act on it.

The aim of all of this is removing ambiguity, not adopting a particular stance. Candidates can adapt to clear rules; they cannot adapt to rules that are different in every recruiter\'s head.

Using AI Ethically: FAQ

What's the clearest definition of "cheating with AI" in a 2026 interview?

Real-time AI assistance during a live interview, without the interviewer's knowledge, where the AI generates the candidate's answer. That's fraud in any reasonable reading. The harder cases sit either side of this — async take-homes, ChatGPT-prepped rehearsal, declared AI use — and require nuance.

Is using ChatGPT for take-home tests cheating?

Depends entirely on what the company specified. If they explicitly prohibited AI assistance and you used it, yes. If they were silent on it and the role you're applying to assumes AI tooling on the job, using AI in the take-home is reasonable — declare it in the submission and explain how. If they explicitly allowed it, no issue. Default: ask.

Should I tell the interviewer I prepped with ChatGPT?

No more than you'd tell them you used flashcards. Preparation is the candidate's responsibility; the tools you used to prepare aren't a reportable variable. The exception: if the interview specifically tests "how would you approach learning X" and you used AI to learn X, mention it — the meta-skill is part of what they're evaluating.

What about using AI for the cover letter?

Acceptable in 2026 if it's used as a starting draft you then edit, not as the final submission. Recruiters can spot pure-AI cover letters reliably (generic phrasing, no project-specific detail) and they signal low effort. AI-assisted-then-edited cover letters are fine and increasingly the norm.

How should companies handle the gray zone?

Be explicit in writing. State whether AI assistance is allowed on take-homes, whether candidates are expected to disclose tooling, and what the consequences of undisclosed real-time use are. Most candidate confusion comes from companies that left the question unaddressed and then enforced inconsistently after the fact.

Published Invalid Date · Updated Invalid Date

Thomas Prommer Technology Executive — CTO/CIO/CTAIO

These salary reports are built on firsthand hiring experience across 20+ years of engineering leadership (adidas, $9B platform, 500+ engineers) and a proprietary network of 200+ executive recruiters and headhunters who share placement data with us directly. As a top-1% expert on institutional investor networks, I've conducted 200+ technical due diligence consultations for PE/VC firms including Blackstone, Bain Capital, and Berenberg — work that requires current, accurate compensation benchmarks across every seniority level. Our team cross-references recruiter data with BLS statistics, job board salary disclosures, and executive compensation surveys to produce ranges you can actually negotiate with.

Profile LinkedIn Newsletter