AI in Video Interviews: What Doctors Need to Know

0Shares

Last Updated: October 2026. AI tools, selection rules and privacy laws are changing quickly. This post is general information, not legal advice. Check your selection body’s candidate rules and seek legal advice for your own situation.

Apps that feed you answers during a video interview are now openly advertised. In this year’s medical selection rounds, candidates using AI in video interviews, and recording those interviews, have become real problems. If you’re a doctor weighing up one of these tools, or you just want to understand what’s changed, this guide explains what the apps do, what employers can detect and where the law draws the line.

TL;DR AI interview apps listen to your video interview and show answers on screen. Detection software can flag them, and selection bodies can disqualify you or refer you to the regulator. Recording the panel without consent may also be a crime. Use AI to prepare, never during the interview.

Below I name some of the apps being marketed to candidates, explain how detection software works and where it falls short, and cover the legal and professional risks. I’ll finish with how to use AI to prepare answers that are unmistakably yours.

Why AI in Video Interviews Has Become a Problem in Medical Selection

Video interviews became standard for doctors’ jobs and training posts during the pandemic, and most selection processes kept them. They save candidates time and travel, especially if you’re rural, regional or overseas. They also opened a door that generative AI has walked straight through.

What I’m hearing this year falls into two groups. First, some trainee doctors have recorded their interviews or been seen looking away from the camera and giving answers that sound scripted. Second, clients tell me, quite sincerely, that they think recording their interview will help them improve. I understand the instinct. As you’ll see below, it can land you in serious trouble.

Selectors are responding. At least one recruitment jurisdiction in Australia has threatened to scrap video interviews altogether. Overseas, NHS England’s medical specialty recruitment has published a formal position. It bans AI-generated responses, virtual assistants, real-time transcription tools and automated scripts during interviews. Candidates who break the rule may be disqualified and referred to their professional or regulatory body. Assessors are trained to look for AI use and may ask you to expand on an answer if they’re suspicious.

It isn’t just medicine. Gartner has predicted that by 2028 one in four candidate profiles will be fake. In its survey of 3,000 job candidates, 6% admitted to interview impersonation fraud, and 40% said they had used AI during their application. Google and McKinsey are among the employers that have brought back in-person interview rounds.

Why Going Fully Face-to-Face Is an Overreaction

I think scrapping video interviews entirely would be an overreaction. It pushes cost, travel and time off work back onto candidates. That hits rural doctors, overseas applicants and anyone with caring responsibilities hardest.

That said, candidates themselves aren’t uniformly against in-person formats. In the same Gartner research, 62% of candidates said they would be more likely to apply to a position if it required an in-person interview — which complicates the “F2F is a burden” framing. The problem with reverting fully to in-person isn’t that candidates hate it; it’s that it still doesn’t solve what video actually broke.

There’s a better answer, and it’s a design problem more than a detection problem. A 2026 paper in npj Digital Medicine argues for AI-resilient interviews rather than a return to in-person formats. That means no unsupervised time with the question before you answer, with interviewers reading the prompt aloud live instead. In that format, candidates who tried to consult AI did worse, because juggling the app added cognitive load and broke the flow of conversation.

A fourth, intermediary option is also quietly gaining traction: local invigilation. Instead of flying candidates interstate, selection bodies ask the candidate to attend a nearby location — a satellite hospital, college office or a trusted invigilator’s practice — where someone supervises the video call in person. It keeps the travel burden low for the candidate while removing the second-screen and earpiece risks that worry panels about remote video.

The AI Interview Apps Being Advertised Right Now

I’m naming these so you recognise them when they turn up in your feed, not as recommendations. Most work the same way: the app listens to the interview audio, transcribes the question and puts a suggested answer on your screen within seconds.

App	What it does during an interview	How it’s marketed
Final Round AI (Interview Copilot)	Listens to the audio and suggests answers built from your CV and the job description. Works with Zoom, Teams, Google Meet and HireVue.	Says it stays invisible when you share your screen
Cluely	Reads audio and on-screen content and overlays AI answers during live calls	Openly pitched as a way to “cheat on everything”
LockedIn AI	Live answer suggestions during interviews	Advertises a “stealth mode”
Interview Coder	Answer help for technical and coding interviews	Built to get past coding interview platforms
Interview Sidekick, GhostPilot	Browser-based tools that listen and suggest answers	Cheap or free tiers
DIY set-up	An AI note-taker transcribes the call while ChatGPT runs on a second device	No product needed

Source: vendor websites and comparison articles, as at September 2026. Features, names and claims change often.

Cluely’s back story says a lot about this market. Its founder was suspended from Columbia University over an earlier interview-cheating tool, then raised US$5.3 million to build a tool to “cheat on everything”. These companies aren’t hiding what they’re for.

Keep one detail in mind for later: nearly all of these apps work by recording and transcribing what the panel says, and usually by sending that audio to an overseas company’s servers.

Not every AI tool is a problem. Mock interview apps and delivery coaches, which give feedback on pace, filler words and eye contact, are legitimate when you use them to practise beforehand. The line is using AI during the real thing.

How Video Interview Software Detects AI Use

No single product catches everything, so employers that take this seriously combine tools. It helps to think of them in three layers.

1. Identity Verification

Before or at the start of the interview, you photograph an identity document and take a “liveness” selfie. This stops someone else sitting the interview for you. Persona, Checkr, Incode, Jumio and Veriff are common providers.

2. Deepfake Detection

These tools analyse voice and video during the call for signs of synthetic media, such as an AI-generated face or cloned voice. Pindrop, Reality Defender and GetReal Security plug into platforms like Zoom and Teams. Pindrop reported that one in six applicants to its own remote roles showed signs of fraud. That’s one company’s hiring data, not a medical figure, but it shows why employers are nervous.

3. AI Assistance Detection

This layer targets the apps in the table above. Talview says it can block apps such as Cluely with a locked-down browser, detect hidden on-screen overlays, use a second camera to scan the room for phones, and analyse audio for AI voice prompts. Sherlock AI (now defunct — a reminder of how fast this market churns) publishes guides on catching Final Round AI and LockedIn AI during live interviews. HireVue and Proctorio offer monitoring and lockdown browsers. Codility, HackerRank and CodeSignal do the same for technical roles.

Layer	What it checks	Example tools	What it means for you
Identity verification	ID document and liveness selfie	Persona, Checkr, Incode, Jumio, Veriff	Have your ID ready and good lighting
Deepfake detection	Synthetic voice or video on the call	Pindrop, Reality Defender, GetReal Security	Avoid filters and virtual backgrounds
AI assistance detection	Overlays, second devices, app activity, gaze and audio patterns	Talview, Sherlock AI, HireVue, Proctorio	One screen, no other apps, phone away

Source: vendor websites, as at September 2026. Detection claims are the vendors’ own and haven’t been independently tested. Naming a tool doesn’t mean any Australian selection body uses it.

Panels don’t need software to notice the tell-tale signs either. Eyes tracking across the screen as if reading, a pause before every answer, and polished but generic answers are all easy to spot. So is an answer that falls apart on the first follow-up question. Experienced interviewers pick these up quickly.

What Detection Misses (and Why That Isn’t a Green Light)

Detection is only partial. A phone propped just out of shot or a helper off camera can slip through. The app makers claim they’re always a step ahead, and the detection vendors claim the reverse. Even Pindrop says technology needs to be paired with well-trained recruiters.

Detection can also get it wrong. Glancing at your own notes, looking away to think, eye movements that don’t match the software’s idea of “normal”, or a laggy connection could all look suspicious to an algorithm. That’s a good reason to make your set-up boringly clean (more on that below).

For doctors, the bigger issue isn’t whether you’d get caught. Honesty sits at the core of the Medical Board of Australia’s code of conduct. A finding that you were dishonest in a selection process can follow you well beyond one job, which is exactly why NHS England refers these cases to regulators. Ahpra has also issued guidance on meeting your professional obligations when using AI. Panels are also judging whether they can trust you with patients, and being caught feeding AI answers to an interview panel answers that question for them.

Recording Your Interview: The Criminal Risk

This is the part most candidates don’t know. Wanting to record your interview so you can review your performance is understandable, but recording a private conversation without consent can be a criminal offence in much of Australia. That includes recording a Zoom or Teams interview with a screen recorder, a phone or an AI note-taker.

What the Law Says

Each state and territory has its own surveillance or listening devices law, and they vary. A “listening device” is broadly defined and can include software that records audio from an online meeting.

In NSW, under section 7 of the Surveillance Devices Act 2007, it is generally an offence to record a private conversation without the consent of all principal parties, even if you’re one of them. The maximum penalty for an individual is 100 penalty units ($11,000) or five years’ imprisonment. There are narrow exceptions, such as recording that is reasonably necessary to protect your lawful interests. Wanting to improve your interview technique is unlikely to count. Sharing or publishing a recording is a separate offence.

Jurisdiction	Main law	Can a participant record without consent?
NSW	Surveillance Devices Act 2007	Generally no, with narrow exceptions
Victoria	Surveillance Devices Act 1999	Generally yes, but sharing or publishing is restricted
Queensland	Invasion of Privacy Act 1971	Generally yes, but sharing or publishing is restricted
South Australia	Surveillance Devices Act 2016	Generally no, with narrow exceptions
Western Australia	Surveillance Devices Act 1998	Generally no, with narrow exceptions
Tasmania	Listening Devices Act 1991	Generally no, with narrow exceptions
ACT	Listening Devices Act 1992	Generally no, with narrow exceptions
Northern Territory	Surveillance Devices Act 2007	Varies; commentators differ, so get advice

Source: state and territory legislation and law firm summaries, as at September 2026. This is a simplified overview, not legal advice. Exceptions and penalties differ between jurisdictions, and where the panel members are located may matter.

Interview panels often sit across more than one state, so a recording that’s arguably lawful where you are may not be lawful where the interviewers are. Don’t gamble on it.

Why AI Apps Make This Worse

An app that listens to and transcribes the panel is, in effect, recording them, and usually passing that audio to a third-party company. You may not think of it as “recording”, but the law looks at what the device does, not what you call it.

Beyond the Criminal Law

Even where recording isn’t a crime, it can still breach the confidentiality terms of a selection process. Sharing interview questions gives some candidates an unfair advantage over others. It also raises professionalism concerns that a college, employer or Ahpra may take seriously. You would also be collecting the interviewers’ personal information without their knowledge.

The simple rule: never record an interview unless the organisation has clearly agreed. If you want to learn from the experience, write detailed notes straight afterwards and ask the selection body for feedback. That’s how my clients get the most improvement anyway.

Your Privacy Rights When Employers Use This Software

The obligations run both ways. If an employer or selection body uses verification or monitoring software on you, you have rights too.

Biometric information is sensitive information under the Privacy Act 1988 (Cth) when it’s used to verify or identify you, so organisations generally need your consent to collect it.
The employee records exemption doesn’t cover applicants who aren’t hired. According to the Office of the Australian Information Commissioner (OAIC), the Australian Privacy Principles still apply to information collected about unsuccessful candidates.
From 10 December 2026, organisations covered by the Privacy Act must explain in their privacy policies when computer programs make, or substantially help make, decisions that significantly affect people.
Victoria has proposed laws to limit AI and biometric surveillance at work, including in recruitment decisions. As at September 2026, these are not yet law.
NSW’s Workplace Surveillance Act 2005 sets notice rules for monitoring employees. Whether it applies to applicants is less clear.
Public health services may also be bound by state privacy and health records laws.

It’s reasonable to ask what’s being collected and which vendor holds it. Ask where it’s stored, how long it’s kept, and whether a human reviews anything the software flags. A good organisation will have ready answers.

How to Avoid Being Wrongly Flagged

Honest candidates can look suspicious on camera without meaning to. A clean set-up removes most of that risk.

Use one screen. Close every other app and turn off notifications.
Put your phone out of reach, face down, on silent.
Skip virtual backgrounds and filters. Use a plain, real background.
Light your face from the front and put the camera at eye level, so you look at the lens rather than off to the side.
Test your connection and use a wired connection if you can.
Ask about notes and adjustments in advance. If you need a reasonable adjustment, such as for disability or hearing, arrange it formally before the day.
If you plan to bring notes, print them out, keep them to one or two pages, and flag it to the panel at the outset — something like “Is it ok if I occasionally refer to some paper notes?” Doing this up front removes any ambiguity if your eyes drift down during an answer.
Think out loud. Saying “Let me take a moment to structure that” explains a pause.
For one-way recorded interviews, read the time limits and retake rules before you start.
If something goes wrong, like a drop-out or a flag, email the recruiter promptly with the facts.

How to Use AI the Right Way to Prepare

I’m not anti-AI. Used before the interview, it’s a genuinely useful coach. The trick is to use it to sharpen your own material, not to replace it.

Predict the questions. Give an AI tool the position description and selection criteria and ask for likely questions.
Run a mock panel. Ask it to play a sceptical interviewer who pushes you with follow-up questions.
Tighten your stories. Paste in your own draft answer and ask where it’s vague, too long or missing a result.
Research the service. Use it to summarise the organisation’s strategic plan and recent news, then check the facts yourself.

Build the content yourself using a structure you can remember under pressure. For “Why this job?” or “Why this specialty?” questions, use VERG™ (Value, Entrustment, Readiness, Gratitude). For behavioural questions, use C-SCARR™ (Competency as signpost, Situation, Complexity, Actions, Results, Reflections). For clinical and ethical scenarios, use RACE™. Answers built this way sound like you, and they hold up when the panel digs deeper, which an AI script never does.

Then rehearse out loud, on camera, to time. Watching recordings of yourself practising is perfectly fine, and it gives you most of what candidates hope to get from recording the real interview.

Prepare Answers That Are Unmistakably Yours

The best protection against AI suspicion is a candidate who clearly knows their own story. My Doctor Job Interview Course walks you through VERG™, C-SCARR™ and RACE™ so you can answer any question with confidence, live and unscripted.

Explore the Interview Course →

The Bottom Line

AI interview apps are real, heavily marketed and tempting when you’re anxious about selection. But they’re increasingly detectable, and they’re banned by selection bodies that have taken a position. Many also record the panel in ways that can be unlawful. The career risk far outweighs any short-term help.

Video interviews are worth keeping, for candidates and selectors alike. Doctors who prepare properly, using AI as a coach rather than a crutch, are the best argument for keeping them. If you’d like more interview help, read my RACS SET interview preparation guide, my tips on avoiding interview failure, and the latest medical recruitment trends.

Get One-on-One Interview Coaching

Have a video interview coming up? Work with the Career Doctors on a realistic mock interview and specific feedback, so you walk in, or log in, ready.

Explore Coaching →

AI in Video Interviews: What Doctors Need to Know

Why AI in Video Interviews Has Become a Problem in Medical Selection

Why Going Fully Face-to-Face Is an Overreaction

The AI Interview Apps Being Advertised Right Now