Should You Tell the Interviewer You Are Using AI?

“Should I tell them I use AI?” doesn’t have a clean yes or no, and anyone who hands you one is skipping the part that actually decides it. Two things settle it: what the AI is doing, and whether anything is being captured. Once those are clear, the disclosure call usually answers itself.

Below you’ll find the common situations, scripts you can copy and use today, and the recording-consent basics that catch people off guard on remote interviews. None of it requires Hedy. And where the law is genuinely murky, we say so rather than pretend it’s settled.

What’s the rule of thumb for disclosing AI?

Run it through two questions. First, what does the AI do? Prepping, researching, and practicing beforehand is broadly fine, and so is a tool quietly taking notes while you talk. AI generating answers you then read aloud as your own, live, is the line almost everyone calls deceptive. Second, is audio being captured? The moment a tool records, consent law kicks in, however innocent the intent.

The real test is simple: are you representing your own capability, or fabricating capability you don’t have? Built In’s recruiter roundup lands in the same place, noting that AI-generated answers are still “fabricated by AI” and that candidates who lean on them tend to flop once they’re on the job (Built In). Prep sharpens a skill you already have. Live answer-feeding fakes one you don’t have in the room. That’s also why honest disclosure rescues the first case and nothing rescues the second. We go deep on the ethics in our piece on whether using AI in interviews counts as cheating. Short version: the deception is what loses people the job, not the AI.

How common is using AI in interviews now?

Common enough that owning up to your own AI use rarely lands as a confession. In the Resume Genius 2026 Job Seeker Insights Report, based on 1,000 active U.S. job seekers, 22% said they had used AI in real time during a live interview, and about 78% use AI somewhere in their job search or would consider it (Resume Genius). A 2025 Blind survey of 3,617 verified professionals found 20% admitted to secretly using AI during interviews, and 55% agreed it “has become the new norm” (Blind, via StartupHub).

And it cuts both ways. Fast Company reported that 21% of U.S. organizations already use generative AI to run at least their first-round interviews (Fast Company). There’s a decent chance your interviewer has an AI notetaker going on their end too. When tools are on both sides of the call, naming yours reads as professional courtesy, not a guilty plea.

When does disclosure actually matter?

It matters most the second something starts capturing audio. It also matters when AI is writing the answers you pass off as your own, or when an interviewer just asks you outright. It barely registers when AI only helped you prep, which most recruiters file alongside any other study resource. Here’s how the common situations map to a move.

Do you have to disclose AI you used to prepare?

Almost never. Using AI to research a company, guess the likely questions, rehearse your stories, or pressure-test a draft answer is no different from hiring a coach or reading a prep book. Built In and other recruiter-facing sources treat preparation as fair game and draw the line at fabricating answers live (Built In). Nobody’s owed an inventory of your study tools.

If you do get asked outright, answer honestly and make clear you were running the tool, not the other way around. The move that lands, as Revarta puts it, is coming across as the thinking human who points AI where they want it, not an operator crossing their fingers that the output is right (Revarta). That reassures people instead of alarming them. For the full prep workflow, including building a story bank and predicting questions, see our guide on preparing for an interview with AI.

When must you disclose recording in an interview?

Any time you capture audio, and the safe move is to ask first. A job interview is usually the interviewer’s meeting, so you showing up with your own notetaker or recorder is the touchier scenario. Announcing it is the bare minimum. Asking permission is better, and it sidesteps the legal question altogether.

The legal backdrop, briefly. U.S. federal law follows one-party consent, so you can record a conversation you’re part of (DMLP). But roughly a dozen states require all-party (sometimes called two-party) consent, where everyone has to agree. The names that come up consistently are California, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania, and Washington, though the exact count and list really do vary by source and by whether the call is over the phone or in person (Wikipedia: Telephone call recording laws). Don’t bother memorizing a number. Check the Justia 50-state survey and your own state. When a remote interview spans two states, the strictest applicable standard wins, so just get everyone’s consent (DMLP).

What about recording laws outside the US?

Europe is stricter, and in several countries covert recording is a crime, not just a privacy headache. Under GDPR, an interview recording counts as “processing of personal data,” so whoever records needs a lawful basis, has to tell the person beforehand, and the person can say no without consequences (Voicit). Then individual countries go further and criminalize secret recording outright.

Germany is the sharp end. Under §201 of the Criminal Code, covertly recording someone’s spoken word can land you up to three years in prison (Cross Channel Lawyers). France’s Article 226-1 of the Penal Code criminalizes recording private speech without the speaker’s consent, with up to one year and a €45,000 fine (FreJun). The UK is easier on you: under RIPA 2000 a participant can record for personal use without telling anyone, but share that recording without consent and you’re likely in breach of the Data Protection Act 2018 (WaveTel). Same lesson everywhere: ask before you capture, and don’t pass the recording around. This is research, not legal advice, and the rules shift by jurisdiction and by how you’re communicating.

What should I say when I use AI?

Steal these and adapt them. They’re short, low-key, and hand the other person an easy out, which is the tone that actually works. For a wider set of recording-consent wording across meeting types, see our scripts for asking permission to record a meeting.

1. Your own AI notetaker (virtual interview):

“Before we start, I like to run an AI notetaker so I can stay present instead of scribbling. It transcribes for my own notes and isn’t shared. Totally fine to turn it off if you’d prefer, just let me know.”

2. If you are asked whether you used AI to prepare:

“Yes. I used it to research your team and pressure-test my own answers. I had it play a skeptical interviewer and poke holes in my examples. The stories are all from my actual experience; the tool just helped me prepare more systematically.”

3. Recorded call (cross-state safe consent):

“Quick heads-up: I’d like to record our conversation so I can review it afterward and follow up accurately. Are you okay with that? Happy not to if you’d rather I didn’t.”

4. Live assistance (the honest non-use, when an interviewer raises AI-cheating concerns):

“I’m not running anything that answers for me. I’d rather you hear how I actually think. If I need a second to pull up a real example, I’ll just take it.”

One thing worth saying out loud: if a prep tool ever feeds you an awkward or flat-out wrong line, own it and recover in your own words rather than dig in. Interviewers are watching how you react when AI fails, and that reaction tells them plenty (Ask a Manager).

Why does honest, low-key framing play better?

Mostly because hiding is the thing that gets you caught, not the AI itself. The detection tech exists (eye-tracking is one example), but the cheapest, most reliable detector is a follow-up question that needs real experience to answer (Built In). Ask a Manager catalogs the tells of someone reading live answers: “eyes darting back and forth like they’re reading,” leaning hard on jargon, and going to pieces on a quick follow-up (Ask a Manager). So hiding loses on the math and on the ethics.

Being open also makes you come across better. People Management cites research that when employers are upfront about AI use, “candidates presented themselves more authentically, resulting in fairer and more accurate hiring outcomes” (People Management). And the norms are moving toward openness on both sides. HiBob CEO Ronni Zehavi says candidates “should always know when AI is used in recruitment,” because “sharing this openly isn’t just about compliance, it’s about respect” (People Management). Only 26% of applicants currently trust AI to evaluate them fairly (Gartner), so being the straightforward one in the room genuinely stands out.

What does Hedy do, and where does it sit on disclosure?

Hedy is a real-time AI conversation coach, not an auto-answer machine. In a virtual interview it points you back at your own material: the achievement on your resume that fits the question, a sharp question to ask back, the result you’d blank on under pressure. It won’t write a script for you to read, and we’ve turned down building “undetectable” or auto-answer features on purpose. That’s not a small product decision. It’s what keeps you on the honest side of the line this whole article is about.

Mapped onto the framework above, Hedy lands where you’d want it. Use it to prep against the job description and your resume, and there’s nothing to disclose. Use it for live coaching in a virtual interview, and it nudges you toward your own experience instead of handing you answers. Audio capture is opt-in and never the default, which is exactly why we say “capture” and not “record”: nothing gets recorded unless you choose it, and when you do, ask first using script 3 above. Processing can run on-device, so sensitive interview content never has to leave your machine. For the full candidate workflow, the job seeker interview tool page walks through it, and you can start free (5 hours a month) before deciding whether Pro at $12.99/month is worth it.

Frequently asked questions

Is it cheating to use AI in a job interview?

Prep with AI is fine. Reading AI-generated answers aloud and presenting them as your own is what nearly all recruiters call cheating (Built In). Research, practice, and feedback are treated like any study tool. The dividing line is representing your real capability versus fabricating capability you don’t have.

Do I have to tell the interviewer I used AI to prepare?

Usually no. Preparation with AI is treated like using a prep book or a coach, and you don’t owe a disclosure of your study methods. If you are asked directly, answer honestly and frame it as you directing the tool, with all your stories drawn from real experience.

Can I use an AI notetaker during a job interview?

Yes, but ask permission first because it captures audio, which triggers consent law. The interview is usually the interviewer’s meeting, so the polite and safe move is to ask rather than just announce. Use a short script that names the tool and offers an easy out.

Is it legal to record a job interview without telling the interviewer?

Sometimes, but it is risky. U.S. federal law allows one-party consent, so you can record a call you are in (DMLP). But roughly a dozen states require all-party consent, and many European countries criminalize covert recording. Because a remote interview can span jurisdictions, just ask.

Which states require all-party consent to record?

The consistently cited core includes California, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania, and Washington, though the list and count differ by source and by phone-versus-in-person rules (Wikipedia). Check the Justia 50-state survey and your specific state rather than relying on a fixed list.

Can interviewers tell when you’re using AI live?

Often, yes. Common tells are answer delays, eyes tracking a second screen, jargon-heavy responses, and a failure to handle quick follow-up questions (Ask a Manager). The most reliable detector is a human follow-up that needs genuine experience to answer (Built In).

Should I ask permission before recording a virtual interview?

Yes. It is the polite default and it moots the entire one-party-versus-all-party question. When participants sit in different states, the strictest standard governs, so universal consent is the safe play (DMLP). And never redistribute a recording without consent, which is where the legal exposure in Europe and the UK actually lives.

Do employers have to disclose when they use AI to interview me?

Increasingly, yes, and the legal obligations run toward employers more than candidates. Ontario requires AI-use disclosure in job postings as of January 2026, and laws like Illinois’ AI Video Interview Act obligate employers (Privacy & Data Security Insights). Many HR leaders also argue disclosure is simple respect (People Management).

How do I tell an interviewer I used AI without sounding like I cheated?

Keep the focus on you directing the tool. Say something like: “I used it to research your team and stress-test my own examples, but the stories are all mine.” That frames you as the thinking human who used AI strategically, not someone relying on it to perform (Revarta).

Is some live AI use ever expected?

Yes, in specific roles. Canva’s engineering team now expects candidates to use AI assistants in technical interviews under a “Transparency Over Detection” philosophy, telling candidates ahead of time (Canva Engineering). When AI is part of the job, employers may want to see you use it. Always confirm the rules first.

Does Hedy feed me answers during an interview?

No. Hedy points you to your own material, an achievement from your resume, a sharp question to ask, a result worth mentioning, rather than writing answers to read aloud. The company has refused to build auto-answer or “undetectable” features by design, which keeps it on the honest, disclosable side of the line.