Caught using AI in an interview: what happens next

What happens if you're caught using AI in an interview?

Usually, you're removed from that hiring process, and sometimes the damage goes further. Deloitte UK says it removes the candidate and may tell their university. Amazon's guidelines warn of disqualification. In the best-known case, reported in March 2025, a student lost his job offers and his university opened disciplinary proceedings.

Most of what's public comes from two kinds of source: employers' own policy pages, and news reports about a few cases. No employer we could find publishes a count of candidates caught. So this page sticks to what employers say they will do and what was reported happening, with a link for each.

What do employer policies say will happen?

The policies that name a consequence agree on the first step: you're out of that hiring process. Some go further. Deloitte UK's integrity guidance for early careers, checked October 22, 2026, says it may report a candidate to their educational institution, and that AI misuse found after hiring can lead to termination of employment.

Amazon's wording comes from internal guidelines quoted by Fortune in June 2025. An Amazon spokesperson told NBC News that candidates must acknowledge they "won't use unauthorized tools" during interviews or assessments. The policy for each employer, and how it changes, is tracked on the page about employer AI interview policies.

Anthropic's candidate AI guidance sets the rule without naming a penalty. Deloitte UK also asks candidates to record a short integrity statement on camera before the online assessment. A candidate who agreed to the rule on video can't easily say later that they didn't know it.

What happened in the best-known case?

The best-documented case is a Columbia University student, Chungin "Roy" Lee, who made it public himself. In February 2025 he posted a video of himself using his own tool, Interview Coder, in an Amazon internship interview. CNBC reported on March 9, 2025, that Amazon and the other companies that had made him offers then rescinded them.

The fallout went past the offers. CNBC reported that Columbia had opened disciplinary proceedings against him, citing documents it had seen, and that YouTube removed the video over a copyright claim by Amazon. Lee later told NBC News the process ended in a yearlong suspension. Columbia declined to comment on individual students.

The accounts differ on one detail. Lee told NBC News that he had turned down Amazon's offer, and that several companies rescinded theirs after he made his AI use public. Either way, none of the offers survived the video.

His case is unusual because he wanted it seen. He told CNBC he never planned to join those companies and wanted to show how easy the format was to game. For a candidate who wanted the job, the same outcome would cost them the job itself.

Why does getting caught cost the offer and not just the round?

An offer is a decision based on what the interviews showed, and these policies treat live AI help as making that evidence unreliable. Deloitte UK's guidance covers discovery at any point, including after you've started work. In the reported case, the problem surfaced after the interviews ended, so the offer was the thing left to withdraw.

That's the rescinded-offer pattern. The interview goes well, and the consequence arrives later. None of the employer pages cited here sets a time limit on acting on what they learn.

It also travels further than one company. Deloitte UK says it may tell your university, and in Lee's case the university acted on its own. A rejection for a weak answer stays between you and one employer. A finding of dishonesty can reach people who were never in the room.

Is using an AI copilot treated like interview fraud?

Employer policies treat live AI help as an integrity breach. The cases that reached criminal courts involved stolen identities instead. In those, people posed as someone else to get hired. The reported cases on this page about candidates who used AI help as themselves ended in lost offers and school discipline, with no criminal charges reported.

The FBI's public service announcement of June 28, 2022, alert number I-062822-PSA, described applicants for remote IT jobs using stolen personal data and deepfakes. In some video interviews, the applicant's lip movements did not match the audio.

On July 24, 2025, the Justice Department announced a 102-month prison sentence for an Arizona woman who helped North Korean IT workers get jobs at 309 US companies. The security firm KnowBe4 wrote on its own blog in July 2024 that it hired a North Korean worker after four video interviews. The background check came back clear because the identity was real and stolen.

These cases are why some employers now check who you are as well as what you say. The changes are collected on the page about how companies counter AI copilots.

What do copilot makers say about being caught?

Some sell on it. Interview Coder's homepage, checked October 22, 2026, calls the product "The No. 1 Undetectable AI For Interviews" and claims "Zero documented cases of users being detected or flagged when using this tool properly." That's the company's own claim. We found no independent test of it, and its founder's case is the best-documented consequence on this page.

A detection claim is a claim about the other side's tools, and those change. Which signals an interviewer can see is covered in can interviewers detect AI copilots. The consequences in the table above don't depend on detection software. Lee's came from a video he posted himself.

The fair part of Lee's argument is about the format. "It doesn't make sense to have an interview format that assumes you don't have the use of AI," he told CNBC. Some employers agree with the premise: Canva's engineering team wrote on June 11, 2025 that it now expects backend, machine learning and frontend candidates to use AI tools in technical interviews, in the open.

What should you do if you're accused of using AI?

Tell the truth, briefly, and in writing if the question came in writing. If you didn't use AI, say so and offer to repeat the question live, on camera or in person. If you did, say that too. A denial that later turns out to be false turns one broken rule into two.

Ask what they observed. A specific question lets you answer the actual worry instead of a guess. Keep the reply to a few sentences and leave out any criticism of their process.

If you use assistive technology for a disability, raise it before the interview. Anthropic's guidance asks candidates to tell their recruiter early about any accommodations. Deloitte UK's page has a section on using AI tools to support a reasonable adjustment. Agreed in advance, it's part of the process.

If the decision goes against you, a short email to ask for feedback after a rejection is still fair. If you want to mention your preparation in a later round, the page on whether to tell the interviewer you used AI has the wording.

How do you use AI without risking any of this?

Use it before and after the call, where the policies above allow or encourage it. Anthropic invites candidates to practice their answers with Claude. Deloitte UK says you may use AI feedback tools to practice for interviews. Both draw the line at the live interview itself, unless the employer says otherwise in writing.

After the call is the other safe window. Write down the questions and your answers while they're fresh, using the free 10-minute debrief template, and mark any answer that had no number or no clear result. The page on whether AI in interviews is cheating sets out before, during and after in one table.

Prept's line is analysis after the call, never during it. If you record your next interview with everyone's consent, the free tier transcribes one interview hour and debriefs it, with no card needed (try one free debrief).