Skip to content
Back to blog

Candidate Fraud Detection in Remote Technical Interviews

By CheatProof.ai Team · Published · Updated

Candidate fraud is deliberate misrepresentation during hiring. In remote technical interviews it can take the form of a proxy interviewer, a fake identity or application, manipulated video, undisclosed AI help, or use of leaked tasks. The appropriate check depends on whether you need to verify identity, corroborate experience, or evaluate who did the work.

Types of candidate fraud

Proxy interviewing and candidate substitution. Someone else answers questions or completes the coding task on the applicant’s behalf. In a bait-and-switch, the person who impressed the panel is different from the person who turns up later. The interview score can be accurate for the person on the call while still failing to evaluate the person you intend to hire.

Fake identities and fake job applicants. An applicant may misstate who they are or use a fabricated work history. Identity and experience are separate checks. A reference can corroborate a responsibility or employment claim, while an identity process confirms that the person corresponds to the verified record. Ask about specific decisions and outcomes so that experience can be evaluated even before references are contacted.

Deepfake video. A video feed can be manipulated to change the apparent speaker. Poor lighting, lag, and dropped frames can also create distracting visual effects. If the call raises a material question, document the issue, offer a second call, and use the established verification process for identity decisions.

Hidden AI assistance. An applicant may use suggestions during a round that prohibits outside help. A desktop assistant’s window can be absent from screen capture. A timing clue or a paste gives interviewers a point to review, and a follow-up can test whether the candidate understands the code. See what to look for when a candidate may be using AI.

Leaked tasks and copied solutions. Someone with advance access to the exact prompt may arrive with a prepared answer. A simple exercise can also produce highly similar solutions from several candidates. Ask every candidate to change a relevant constraint or explain a trade-off. Replace exposed tasks with equivalents of similar difficulty.

Which checks catch which type

Assign each check to a stage and a person on the hiring team. An interviewer can assess coding evidence during a live round, while the team handling an offer can conduct identity and reference checks. Record what the check actually established and who will resolve any remaining discrepancy.

Candidate fraud risks and corresponding hiring checks
RiskCheck in the hiring processFollow-up
Proxy interview or candidate substitutionUse a camera-on round where appropriate and an accessible alternative where needed. Have a different interviewer run a final live round, by video or in person, and check whether the same person attended earlier rounds.Ask the candidate to explain a decision they made in the first round. Route a material identity discrepancy to the team responsible for verification.
False identity or fabricated work historyRun identity verification at the offer stage and reference checks for material experience claims. Use the same process for comparable candidates.Document which claim needs corroboration and give the candidate a way to correct a clerical error or explain a name difference.
Manipulated or synthetic videoArrange a second live conversation if the first call leaves an unresolved discrepancy. Use a separate verification process for high-confidence identity checks.Record the observable issue, such as recurring audio and video mismatch, and test the connection before escalating.
Undisclosed AI assistanceState which tools are allowed. Use follow-up questions, editor activity and, when appropriate, desktop proctoring and AI code comparison.Review the relevant moment and ask the candidate to explain or extend the code. Score the response against the interview rubric.
Leaked task or copied solutionMaintain a bank of equivalent tasks and use a live variant question. Review code similarities alongside how the solution was developed.Check whether ordinary solutions converge on the same code. Ask the candidate to handle a new edge case before deciding what the match means.

For example, a candidate who performs differently in two rounds may have had different tasks, conditions, or time limits. Compare what each round measured before investigating identity or assistance. If a specific work claim conflicts with a reference, investigate that claim directly instead of asking the technical interviewer to make an identity judgment from a video.

Where CheatProof helps — and where it doesn't

CheatProof helps interviewers review how a live-coding solution was produced. The browser IDE records activity such as page departures, returns, and pastes into the editor. AI plagiarism analysis compares the submitted code with AI-written solutions and highlights matches. Similarity can result from common approaches, so use it alongside the candidate’s explanation and the code history.

The macOS desktop app can flag a window that appears on the monitored display but is missing or nearly blank in screen capture. Review when the window appeared and whether code was pasted around the same time. CheatProof cannot inspect a separate device. Recording requires candidate consent. See the guides to Cluely and Interview Coder for examples of what a hidden-window signal can tell you.

CheatProof does not verify identity or detect deepfakes. Keep those decisions in your hiring process or with a separate verification service. Assessment tools also report different signals depending on their settings. If your team uses a separate coding assessment before the live round, see what each HackerRank proctoring mode documents.

A simple anti-fraud checklist for remote hiring

  1. Set the rules before the call. Specify which documentation, editors, and AI tools candidates may use. Explain any recording or desktop monitoring and how candidates can request an accommodation.
  2. Make the test comparable. Give candidates equivalent tasks and a shared rubric. Rotate exposed prompts and preserve a record of which version each candidate saw.
  3. Ask for one live change. Give the candidate a new input or requirement and ask them to predict the result before editing. Record their reasoning and the test they choose.
  4. Record observations precisely. Write down the timestamp and what the tool or interviewer observed. Keep the candidate’s explanation in the same review record.
  5. Have a second person review material discrepancies. Give them the task rules, the relevant evidence, and the candidate’s response. Ask them to assess the same rubric independently.
  6. Check continuity between rounds. Share relevant interview notes with the next interviewer. If a discrepancy remains, arrange a focused live round and let the candidate address it.
  7. Verify identity and references at the assigned stage. Use a consistent offer-stage identity process and a clear owner for reference checks. Keep verification documents in the designated hiring system.

Set up one live interview to see how the coding evidence fits your review process. New organizations with a verified work email can try one free.

FAQ

What is candidate fraud?

Candidate fraud is deliberate misrepresentation during hiring. It includes false identities or credentials, another person taking an interview, manipulated video, undisclosed assistance that violates the interview rules, and using leaked answers as if they were original work.

How do you spot a fake job applicant?

Verify specific claims at the appropriate stage. Use structured questions about prior work, reference checks, and a consistent offer-stage identity verification process. If something does not match, record the discrepancy and give the applicant a chance to resolve it before you make a decision.

Can a candidate use someone else in a video interview?

Yes. A proxy can attend an interview or complete an exercise on someone’s behalf. A second live round with a different interviewer can help check continuity across rounds. Use your separate identity verification process for a formal identity decision.

Does CheatProof detect deepfakes?

No. CheatProof does not detect deepfakes or verify identity. Its live-coding tools flag capture-hidden windows on a monitored desktop and surface editor activity and code similarities for an interviewer to review.