Proxy interviews: how to spot and stop them

Remote hiring made it easier to interview candidates anywhere in India. It also made it easier for someone else to sit the interview. Proxy interviews are not new, but video calls, cheap earpieces and AI face filters have made them harder to spot. This guide explains how they work, the warning signs, and a set of process checks any hiring team can put in place.

What a proxy interview is

A proxy interview is when someone other than the real applicant takes part in a hiring round on their behalf, or feeds them answers live during it. The stand-in may sit the whole interview, only the technical round, or stay off camera and supply answers through an earpiece or a hidden screen. The real applicant then joins the company and does the job.

It is different from a candidate using an AI copilot, although the two overlap. With a copilot, the help comes from software. With a proxy, it comes from a person. Some tools now offer both: LockedIn AI markets a “Duo” feature that brings a human helper into the interview, and says “the interviewer sees only you and your normal screen share”.

How proxy rings operate

SpringVerify, a background verification firm, describes three common versions:

1. Full substitution. A different person appears on video, using the applicant’s name. This relies on nobody comparing the interview with the applicant’s identity documents.
2. Live answer feeding. The real applicant is on camera, while a proxy off camera listens and feeds technical answers through an earpiece or a second screen.
3. AI-assisted impersonation. A face filter or deepfake places the applicant’s face over the proxy’s video. SpringVerify describes this as still emerging but technically feasible with consumer tools.

SpringVerify also reports that some services in India charge ₹20,000 to ₹50,000 to provide a professional interviewer who impersonates the candidate. That is one firm’s report, but it shows this is an organised trade, not only friends helping friends.

Real cases do surface. In a case reported in April 2025, a software engineer who had joined Infosys in Bengaluru was found to have sent a friend to his virtual interview. Colleagues noticed that his communication at work did not match the interview. HR compared images from the virtual interview with the employee, found they were different people, and the company filed a police complaint. Police booked him for cheating and cheating by personation.

The problem is not limited to India. In a Checkr survey of 3,000 US hiring managers published in September 2025, 35% said someone other than the listed applicant had taken part in a virtual interview. As early as June 2022, the FBI’s Internet Crime Complaint Center warned of complaints about deepfakes and stolen personal data being used to apply for remote IT, programming and database roles.

Warning signs

No single sign proves anything, but several together are worth acting on.

• Lips and voice out of sync. The FBI’s 2022 warning noted that in deepfake interviews, “the actions and lip movement of the person seen interviewed on-camera do not completely coordinate with the audio”, and that coughing or sneezing did not line up with the video.
• Camera “problems”. Repeated requests to stay audio-only, a frozen or very low-quality picture, or video that drops whenever you ask a hard question.
• Answers that don’t match the CV. Detailed knowledge of technologies that aren’t on the CV, and vagueness about projects that are.
• A different voice or manner in later rounds. Fluent and confident in the technical round, hesitant in the HR round, or the other way round.
• Eyes and pauses. Repeated glances to one side and a pause before each answer, as if waiting for someone else.
• Background noise. Murmuring, typing when the candidate isn’t typing, or a muted microphone at odd moments.
• A drop after joining. As in the Infosys case, the clearest sign often comes in the first weeks of work, when ability falls well short of the interview.

Be careful with appearance. People look different on a webcam than in person. Weight changes, new glasses, a haircut or poor lighting are not evidence of fraud.

Process checks that make a stand-in easier to catch

Most proxy attempts rely on a gap in the process: nobody compares faces, different people run each round, or the identity check happens weeks later. Close those gaps.

• Check ID at the start of the call. Ask the candidate to hold a government photo ID up to the camera and compare it with their face. Record that you did this on the scorecard. This is a step for the interviewer. It is not automated by any tool described here.
• Keep one person across rounds. Have at least one interviewer, or a recruiter, appear in every round so someone can notice a change in face, voice or manner.
• Make the technical round live. Ask the candidate to write, run and explain code on a shared code pad. Then change the problem. A person being fed answers struggles with fast, specific follow-ups.
• Compare recordings between rounds. Before an offer, watch a short clip from each round, one after the other.
• Check identity again at joining. At onboarding, in person or on video, compare the new employee with the interview recording and the ID seen in the interview.
• Set a first-week task. SpringVerify suggests a technical task in the first week that mirrors the interview’s difficulty. A large gap is worth a conversation.

For candidates who get live help from AI tools instead of a person, see our guide to the signs a candidate is using AI.

How recordings help

The Infosys case was caught by comparing an image of the interview with the person who turned up. That works far better when every round is recorded as a matter of routine, not only when someone happens to capture an image.

Eagle Vision records every live interview: the video call with the candidate on camera, the shared code pad, the live flags and the integrity score. Run your live interviews in fullscreen lockdown and you have a reference recording of the face and voice from each round. Recordings are watched in the app by the interviewer who hosted the round or by an admin, so an admin can open earlier rounds before the next one. Your organisation’s owner chooses whether recordings are kept for 30, 45 or 60 days. Within that period, the hiring team can rewatch a recording in the app for an identity check at onboarding. Recordings can’t be downloaded or shared by link, and notice periods often run longer than 60 days, so make sure someone who sat in the interviews or reviewed the recordings also joins the onboarding check.

The lockdown also helps with the live-feeding version of the fraud. If answers are arriving through another window, switching tabs, moving the cursor to another screen, and copying or pasting text are flagged as they happen. Each flag is timestamped, so from the report you can jump straight to that moment in the recording.

To be clear about the limits: Eagle Vision does not verify identity automatically and has no face detection or deepfake detection. It will not tell you that the person on camera is the applicant. It gives your interviewers a controlled session and a full recording, so the checks above are easier to carry out and to evidence afterwards. An earpiece or a helper out of the camera’s view will not raise a flag, which is why the live, fast-moving technical round matters so much.

A fair process and candidate consent

Most candidates are honest, and the checks should not treat them as suspects.

• Tell candidates in advance that you will ask to see photo ID on camera and that interviews are recorded. Explain how recordings are stored and for how long, and link to your privacy policy.
• Apply the same checks to everyone, so no candidate is singled out because of how they look or sound.
• Look only at what you need. Ask candidates to show the ID on camera rather than send a copy, unless you genuinely need one.
• Investigate before you act. If something doesn’t add up, invite the candidate to a short second conversation, on video with a different interviewer or in person.
• Document the decision, including what you saw and when.

Our guide on how to prevent cheating in online interviews covers how to communicate these rules.