How to spot an AI-generated or fake resume, and why only one of them matters

An AI-generated resume and a fake resume are different problems. Most candidates now use AI to draft or polish their resumes, and no tool can reliably tell you who wrote the sentences, so "AI-written" is neither detectable nor disqualifying. A fake resume is one whose facts are false: invented employers, borrowed skills, unverifiable credentials, or a whole identity. To spot one, ignore the writing style and check the claims: search for each employer yourself, compare the resume with the candidate's profiles, test the contact details, look for hidden text, and turn the three claims that matter most into phone-screen questions.

This page covers the resume itself, as of October 2026. Candidates who apply in bulk with bots, or who are not real people at all, are covered in fake job applicants, and hidden instructions aimed at AI screeners in resume prompt injection.

Is an AI-written resume a problem?

On its own, no. In Gartner's survey of 3,290 candidates in late 2024, 39% used AI during the application process, and 54% of those used it to generate resume text. By October 2026, Greenhouse found 77% of US job seekers had used AI in their search in the past year, and 39% had rewritten an application so it would not be mistaken for AI. Screening out "AI-sounding" resumes therefore filters on style while the content goes unchecked.

The content is where the trouble is. In the same October 2026 Greenhouse survey, 62% of US job seekers had tried tactics to get past automated screening, from using AI to work out what the system looks for (54%) to adding skills or keywords they do not have (40%). In Greenhouse's July 2025 survey, 32% said they had claimed AI skills they do not actually have. Equifax found that 71% of HR professionals had encountered fabricated or misleading candidate information, most often in employment history, education and credentials (August 2025, a vendor survey of 254 HR professionals).

Can AI detectors tell if a resume was written by ChatGPT?

Not reliably enough to act on. When OpenAI released its AI-text classifier in January 2023, it reported that the tool correctly identified 26% of AI-written text and wrongly labeled human-written text as AI-written 9% of the time; OpenAI withdrew it on July 20, 2023, because of its low accuracy. A study by Liang and colleagues published in Patterns (2023) found that widely used detectors consistently misclassified writing by non-native English speakers as AI-generated. Resumes make it worse: they are short, formulaic and full of standard phrases, the conditions under which detectors do least well.

Recruiters are not confident either. In HireRight's 2025 benchmark survey, 34% of North American respondents said they had little to no confidence they could tell when candidates used generative AI on applications or resumes. That is the right instinct: stop trying to detect authorship and verify facts.

What are the signs of a fake resume?

None of these proves anything alone. Two or three together justify verifying before you spend interview time.

What is not a sign of a fake resume?

• Polished, generic phrasing. That is what resume templates, coaches and AI all produce.
• Particular punctuation or vocabulary. Style "tells" circulate online and change monthly; they say nothing about whether the facts are true.
• A gap, a short tenure or a career change. These deserve a question, not suspicion; see how to probe resume gaps.
• A school or employer in another country. Treating someone's origin as a fraud signal is unfair and creates discrimination risk; verify the claim like any other.

How can you check a resume in ten minutes?

1. Choose the three claims that matter most for this role: usually the most recent employer, the key skill and any required credential.
2. Search each employer yourself. Does it exist, did it exist at the dates claimed, and is it the size the title implies? A "VP" at a three-person company is possible, but worth a question.
3. Compare the profiles. Dates, titles and photo across the resume, professional profile and any portfolio.
4. Test the contact details. Note whether the number works and search your system for duplicates.
5. Reveal hidden text. A plain-text paste takes seconds and shows anything in white or tiny type.
6. Check one credential with the issuer's own registry; professional license verification lists where to look.
7. Write the three claims into your screen as questions, below.

Which phone-screen questions test a resume?

Ask about the claims you chose, then follow up twice. Real experience gets more specific; borrowed experience gets vaguer.

"At [employer], what did the team look like when you joined, and who
did you report to?"
"You list [skill]. What's the last thing you built or fixed with it,
and what went wrong first?"
"How was [number on the resume] measured, and who would disagree
with it?"
"Your [certification]: when did you get it, and what did you study
for it that you've used since?"

If a candidate cannot discuss their own resume in detail, that is the finding, whoever wrote the document. More questions built on this principle are in interview questions that expose AI-assisted answers.

What about AI-generated photos and profiles?

The resume is often only one piece of a constructed identity. Microsoft reported in June 2025 that North Korean IT workers use AI tools such as Faceswap to move their pictures onto stolen employment and identity documents and to make their photos look more professional. In Resume Genius's early-2025 survey of 1,000 US hiring managers, 33% said they had spotted AI-generated user profiles on professional networks or social media (a vendor survey). Practical checks:

• Does the profile have a history (posts, recommendations, connections at the claimed employers) or did it appear recently, fully formed?
• Does the same photo appear elsewhere under another name? A reverse image search takes a minute.
• Do people at the claimed employers know the candidate? A reference you find yourself answers this better than one the candidate supplies.
• Does the person on the video screen look like the photo, and stay the same person across rounds?

When should you verify formally?

Before an offer, for every finalist, not only the ones who seem suspicious. Confirm employment through contacts you find yourself (how to verify employment history), degrees with the school or a verification service (how to verify education credentials), and references by live conversation (how to spot fake references). If you use a background screening company, the Fair Credit Reporting Act's consent and notice rules apply; see the FCRA background check process.

What should you do when a resume doesn't check out?

• Separate exaggeration from fabrication. A rounded-up date or a generous title is a conversation; an employer that does not exist is not. How to handle a candidate who lies in an interview covers the difference.
• Ask before you conclude. "We couldn't find a record of [company]. Can you point us to someone there who can confirm your role?" Companies close, merge and rename; a real candidate can usually help.
• Escalate identity concerns separately. If the person, not just the history, may be invented, follow how to spot candidate identity fraud.
• Record facts. "Employer X not found in registry or online; candidate did not provide a contact when asked on [date]" is a note you can defend. "AI-generated resume" is not.