# Can ATS Systems Detect AI Generated Resumes?

> Can ATS systems detect AI generated resumes? What screening software checks, why AI detectors misfire, what recruiters notice, and how to use AI safely.

Canonical: https://tryelip.ai/blog/can-ats-detect-ai-resumes
Last updated: 2026-11-20

Published 2026-11-20.

Can ATS systems detect AI generated resumes? The applicant tracking systems whose documentation we read do not describe any check for whether a person or an AI wrote your resume. They parse it, apply the employer's screening rules, and in some cases grade how well your qualifications match the job. The detecting is done by people, and in surveys a minority of hiring managers say they would reject an AI-written resume. Standalone AI text detectors exist, but published tests find them unreliable. The risk sits in what the AI writes for you: generic lines, claims you cannot back up, and hidden text.

## What an ATS checks

An applicant tracking system is the software employers use to collect and manage applications. Its main jobs, as described in vendor documentation, are reading your file, applying the employer's rules and sorting applicants.

- **Parsing.** Greenhouse's help page on [unsuccessful resume parses](https://support.greenhouse.io/hc/en-us/articles/200989175-Unsuccessful-resume-parse) lists what breaks it: files over 2.5MB, resumes uploaded as images, graphics, tables, columns, and contact details in a header or text box.
- **Screening rules.** Greenhouse's [auto-reject rule](https://support.greenhouse.io/hc/en-us/articles/203105595-Application-Rules-Overview) "uses custom application questions to screen and reject candidates based on their responses."
- **Matching.** Workday's HiredScore feature gives candidates [grades from A to D](https://doc.workday.com/hiredscore/en-us/workday-hiredscore/recruiter-productivity-/reference--candidate-grades.html) based on "how well their resumes match the skills and qualifications specified in the job description." Greenhouse's [Talent Matching](https://support.greenhouse.io/hc/en-us/articles/41396009937307-Talent-Matching) sorts candidates into categories from "Strong match" to "Limited match" against criteria the employer sets.
- **Identity.** Greenhouse's Real Talent offers an optional identity check through CLEAR, a selfie and ID check, after which recruiters see a "Verified" status, according to [Greenhouse's own description](https://www.greenhouse.com/blog/building-trust-through-transparency-with-real-talent).

None of these pages describes scoring a resume for AI authorship. Greenhouse states that Talent Matching "does not make hiring decisions", and its [AI principles](https://www.greenhouse.com/ai-principles) say AI "is never the final decision-maker."

An employer could still paste your resume into a separate AI detector. That is a choice made by a person, outside the ATS, and the tools have a poor record.

## Why AI text detectors are unreliable

Two published studies are worth knowing:

- An international team of researchers led by Debora Weber-Wulff tested 14 detection tools, including two commercial ones. They concluded the tools were ["neither accurate nor reliable"](https://arxiv.org/abs/2306.15666) and tended to label text as human-written.
- Researchers led by Weixin Liang and James Zou found that GPT detectors ["consistently misclassify non-native English writing samples as AI-generated"](https://arxiv.org/abs/2304.02819) while correctly identifying native writing samples. Their paper is titled "GPT detectors are biased against non-native English writers."

A resume is also a poor fit for these tools. It is short, it follows a fixed format, and most people use the same stock verbs. A detector score on a resume tells a recruiter very little, and a careful employer knows that.

## What hiring managers notice instead

People do try to spot AI writing, with mixed results:

- In a [TopResume survey](https://topresume.com/career-advice/ai-in-hiring-survey) of 600 US hiring managers in May 2025, 33.5% correctly picked out the AI-generated resumes from a set of four. 19.6% said they would reject a candidate with an AI-generated resume or cover letter, and 52% said using AI for proofreading or drafting support is acceptable.
- In Greenhouse's [2025 AI in Hiring Report](https://www.greenhouse.com/newsroom/an-ai-trust-crisis-70-of-hiring-managers-trust-ai-to-make-faster-and-better-hiring-decisions-only-8-of-job-seekers-call-it-fair), 65% of US hiring managers had caught applicants using AI deceptively. 22% had found prompt injections hidden in resumes, and 32% had caught candidates reading from AI-generated scripts.

The common thread is deception. A resume drafted with AI and checked by you is acceptable to most of these managers. A resume with invented skills, hidden instructions or interview answers read off a screen is what gets people caught.

Job seekers have noticed. In Greenhouse's [2026 candidate survey](https://www.greenhouse.com/newsroom/6-in-10-job-seekers-are-rethinking-how-they-write-to-avoid-sounding-like-ai-even-as-most-use-it-in-their-search), 77% of US job seekers had used AI in their search in the past year, and 60% had changed or considered changing how they write to avoid sounding like AI.

## Signs a resume reads as AI-written

These are the patterns that make a reader suspicious, whoever wrote the text:

1. Bullets with no numbers, names or results, such as "Drove impactful cross-functional initiatives."
2. The same three adjectives in the summary that appear in thousands of other summaries: results-driven, passionate, dynamic.
3. Every bullet the same length and shape.
4. Skills that match the posting word for word but never appear in the experience section.
5. A summary that repeats the job description back to the employer.
6. Claims that fall apart at the first follow-up question in an interview.

Here is the difference in practice:

- AI-generic: Spearheaded innovative strategies to enhance customer satisfaction and drive operational excellence.
- Specific: Cut average support reply time from 9 hours to 3 by moving 40% of tickets to a shared macro library.
- AI-generic: Leveraged data-driven insights to optimize marketing performance.
- Specific: Rebuilt the weekly paid search report in Looker; the team paused 6 campaigns that were losing money within a month.

The specific versions are harder to fake and easier to talk about. They also match better, because they name tools and results the posting asks for.

## Where AI helps on a resume, and where it hurts

AI is useful for the parts of resume writing that are mechanical:

- Sorting a job description into must-have and nice-to-have requirements.
- Spotting requirements your resume does not mention yet.
- Shortening long bullets and fixing grammar.
- Suggesting a clearer order for your sections.
- Checking the same date format and tense across every job.

It hurts when it does the parts that need your knowledge:

- Writing achievements from nothing. It will produce plausible results you never had.
- Inflating numbers or scope to sound stronger.
- Adding skills from the posting that you have only heard of.
- Rewriting everything in the same polished, empty voice.
- Writing a summary that could belong to anyone applying for the role.

A simple rule: let AI handle form, and keep the facts yours.

## How to use AI on your resume without the risk

1. **Start from a facts sheet.** Write down your roles, dates, tools, results with real numbers, and certifications. This is the only material the AI may use.
2. **Ask for structure, not invention.** Have the AI reorder and tighten your bullets for a specific job. Tell it not to add anything.
3. **Replace every generic phrase.** For each bullet, ask yourself what you did, with what, and what changed. Put that in.
4. **Read it out loud.** If a line does not sound like something you would say in an interview, rewrite it.
5. **Remove anything hidden.** No white text, no tiny fonts, no instructions to an AI reviewer.
6. **Check every claim.** If you could not explain it with an example in an interview, cut it.

A prompt you can copy:

> Here is my facts sheet and a job description.
> Rewrite my experience bullets for this job.
> Rules:
> - Use only facts from my sheet. Do not add skills, tools, numbers or titles.
> - Keep my own words where they are already specific.
> - Each bullet: what I did, the tool or method, and the result with its number.
> - Avoid these words: spearheaded, leveraged, dynamic, results-driven, passionate.
> - List separately any requirement in the job that my sheet does not support.
> 
> Facts sheet:
> [paste]
> 
> Job description:
> [paste]

Then do the final edit yourself. Our guide to a [tailored resume](https://tryelip.ai/blog/tailored-resume) has a full worked example, and the [AI job search guide](https://tryelip.ai/blog/ai-job-search) covers using AI across the rest of the search.

## If an employer runs a detector anyway

Assume some employers do. A careful recruiter will treat a detector score on a resume as weak evidence, given the research above, but you cannot know who is careful. The defense is the same either way. A resume full of your own specifics, such as named tools, real numbers, client types and team sizes, reads as written by someone who did the work, and it gives an interviewer plenty to ask about. If a recruiter raises a detector result, offer to walk through any bullet in detail. That conversation settles the question faster than any score.

## A last check before you submit

1. Every bullet has a number, a named tool or a named result.
2. You can tell a short story about each of your top three bullets.
3. No skill appears that you would struggle to discuss for two minutes.
4. The summary mentions this role or this kind of company.
5. Nothing on the page is hidden, tiny or the same color as the background.
6. The wording sounds like you when you read it out loud.

## If an employer asks whether you used AI

An interviewer or an application form may ask you directly. Answer honestly and briefly. Using AI as an editing tool is common: in Greenhouse's 2026 survey, 77% of US job seekers had used AI in their search. What matters to the employer is that the content is yours and true. You can adapt this:

> Yes. I used an AI assistant to tighten the wording and check the layout. The experience, results and numbers are my own, and I am happy to walk you through any of them.

Then be ready to do exactly that. Pick the three bullets you are most likely to be asked about and prepare a two-minute story for each: the situation, what you did, the tool or method, and the result.

## Cover letters and application answers

The same logic applies beyond the resume. Free-text answers on application forms and cover letters are where generic AI writing shows most, because the employer asked you a specific question. If a form asks why you want this job, the answer should name something about this company that you found yourself. Use AI to shorten a draft or fix grammar, then make sure at least one sentence in every answer could only have been written by you.

## What gets resumes rejected

If your resume is not getting responses, AI detection is an unlikely cause. The documented causes are more ordinary: screening question rules, missing required qualifications, layouts that do not parse, and applying after a shortlist is full. Our list of [common ATS resume rejection reasons](https://tryelip.ai/blog/ats-resume-rejection-reasons) goes through each one with a fix, and [how AI resume screening works](https://tryelip.ai/blog/ai-resume-screening) explains the grading step.

For the bigger picture, read [what an AI career agent does](https://tryelip.ai/guides/ai-career-agent), and the numbers in our [job search statistics](https://tryelip.ai/research/job-search-statistics).

## Finding your next role

[Meet Elip](https://tryelip.ai/), an AI career agent. Add your resume and say what you want. Each morning it brings you a few jobs that fit, each with the reason. It rewrites your resume for each role, with your approval. It is free for candidates.

## Sources

- Greenhouse Support, "Unsuccessful resume parse": [https://support.greenhouse.io/hc/en-us/articles/200989175-Unsuccessful-resume-parse](https://support.greenhouse.io/hc/en-us/articles/200989175-Unsuccessful-resume-parse)
- Greenhouse Support, "Application Rules Overview": [https://support.greenhouse.io/hc/en-us/articles/203105595-Application-Rules-Overview](https://support.greenhouse.io/hc/en-us/articles/203105595-Application-Rules-Overview)
- Greenhouse Support, "Talent Matching": [https://support.greenhouse.io/hc/en-us/articles/41396009937307-Talent-Matching](https://support.greenhouse.io/hc/en-us/articles/41396009937307-Talent-Matching)
- Greenhouse, "Building trust through transparency: How we're shaping Real Talent": [https://www.greenhouse.com/blog/building-trust-through-transparency-with-real-talent](https://www.greenhouse.com/blog/building-trust-through-transparency-with-real-talent)
- Greenhouse, "Our AI principles": [https://www.greenhouse.com/ai-principles](https://www.greenhouse.com/ai-principles)
- Workday HiredScore documentation, "Candidate grades": [https://doc.workday.com/hiredscore/en-us/workday-hiredscore/recruiter-productivity-/reference--candidate-grades.html](https://doc.workday.com/hiredscore/en-us/workday-hiredscore/recruiter-productivity-/reference--candidate-grades.html)
- Weber-Wulff et al., "Testing of Detection Tools for AI-Generated Text" (arXiv 2306.15666): [https://arxiv.org/abs/2306.15666](https://arxiv.org/abs/2306.15666)
- Liang, Yuksekgonul, Mao, Wu, Zou, "GPT detectors are biased against non-native English writers" (arXiv 2304.02819): [https://arxiv.org/abs/2304.02819](https://arxiv.org/abs/2304.02819)
- TopResume, "Survey: Where Employers Draw the Line on the Use of AI in Hiring" (June 3, 2025): [https://topresume.com/career-advice/ai-in-hiring-survey](https://topresume.com/career-advice/ai-in-hiring-survey)
- Greenhouse, 2025 AI in Hiring Report release (November 19, 2025): [https://www.greenhouse.com/newsroom/an-ai-trust-crisis-70-of-hiring-managers-trust-ai-to-make-faster-and-better-hiring-decisions-only-8-of-job-seekers-call-it-fair](https://www.greenhouse.com/newsroom/an-ai-trust-crisis-70-of-hiring-managers-trust-ai-to-make-faster-and-better-hiring-decisions-only-8-of-job-seekers-call-it-fair)
- Greenhouse, 2026 candidate survey release (October 20, 2026): [https://www.greenhouse.com/newsroom/6-in-10-job-seekers-are-rethinking-how-they-write-to-avoid-sounding-like-ai-even-as-most-use-it-in-their-search](https://www.greenhouse.com/newsroom/6-in-10-job-seekers-are-rethinking-how-they-write-to-avoid-sounding-like-ai-even-as-most-use-it-in-their-search)

## Frequently asked questions

### Can ATS detect AI-written resumes?

The ATS documentation we read, from Greenhouse and Workday, describes parsing, screening rules, qualification matching and optional identity checks, with no check on whether AI wrote the text. An employer could run your resume through a separate AI detector, but published tests have found those tools neither accurate nor reliable.

### Do recruiters care if you use AI for your resume?

Some do. In a TopResume survey of 600 US hiring managers, 19.6% said they would reject a candidate with an AI-generated resume or cover letter, while 52% said AI for proofreading or drafting support is acceptable. What managers report catching most is deceptive use, such as hidden prompts or invented claims.

### Can employers tell if ChatGPT wrote my resume?

Sometimes. In TopResume's test, 33.5% of hiring managers correctly identified the AI-generated resumes. Generic phrasing, bullets with no numbers, and claims that you cannot explain in an interview are the giveaways. A resume built from your own specific facts is hard to tell apart from one you wrote alone.

### Will an AI detector flag my resume?

It might, whether or not you used AI. A study of 14 detection tools found them neither accurate nor reliable, and another study found GPT detectors consistently misclassify writing by non-native English speakers as AI-generated. Short, formulaic text like a resume is a weak input for these tools.

### Is it OK to use AI to write a resume?

Yes, for drafting and editing, as long as every claim is true and you have checked every line. Use your own facts as the only input, replace generic phrases with specifics, and never add skills you do not have. In Greenhouse's 2026 survey, 40% of US job seekers had tried adding skills or keywords they lack.

### Should I hide keywords in white text to pass the ATS?

No. Hidden text is one of the things hiring managers report catching: in Greenhouse's 2025 report, 22% of US hiring managers had found prompt injections hidden in resumes. Anything a reader cannot see should come out. Put real, visible keywords where they honestly describe your work.

## Related

- [All posts](https://tryelip.ai/blog)
