Survey data on how candidates use AI to write resumes, how often employers can tell, and what — if anything — they do about it.
AI-assisted resume writing went from a curiosity in 2023 to a near-default workflow by 2026. The honest empirical questions — how often candidates use AI, how often employers notice, and whether AI-written content actually performs worse — are partially answered by survey data and partially still emerging. Here's what the published research and public industry data say so far.
ResumeBuilder.com's recurring surveys of US job seekers consistently find that roughly half of recent applicants have used AI tools in some part of their resume-writing process — drafting bullets, generating summaries, or rewording entire sections.
Sources: [1]
The self-reported detection rate is high, though it conflates "I can sometimes tell" with "I can reliably tell." Independent text-classification research (OpenAI's own and academic detectors) generally shows much lower accuracy at distinguishing AI from human writing in short, edited text.
Sources: [2]
As of 2026, Workday, Greenhouse, Lever, and Taleo have not added explicit AI-detection features visible to recruiters. Recruiter-tooling vendors (Eightfold, HireEZ, Gem) have publicly discussed experimentation in this area but have not shipped general-availability detection features.
Sources: [3]
Greenhouse has shared industry data showing the volume of resumes with detectable AI signatures roughly doubled in twelve months, with the largest growth in tech and marketing roles.
Sources: [4]
Industry data on application-to-interview conversion has not shown a clear penalty for AI-assisted resumes in aggregate. Human-review studies suggest that AI-generated text that has been edited performs comparably to fully human-written resumes; un-edited AI output (e.g., resumes that read as a clean ChatGPT export) performs worse.
Sources: [5]
This is an emerging research area. Most data comes from industry surveys (ResumeBuilder.com, Capterra) and reports published by ATS / recruiter-tooling vendors. Where survey respondents self-report ability to detect AI content, we flag that as self-reported rather than measured. Independent academic research on text classification is ongoing.
Used as an editor, AI is a clear net positive for resume-writing — comparable or better quality, dramatically less time. Used as a writer with no editing, AI produces resumes that perform measurably worse on human review. The reasonable workflow is: write specifics yourself, use AI to polish phrasing and tighten bullets, then edit the AI output back into your own voice. Treating AI as a co-author rather than a ghostwriter is the version that ships well.
The famous "6 seconds" claim, where it comes from, what the eye-tracking data actually measured, and how to write a resume for the way recruiters really read.
How many resumes really get rejected before a human sees them, who's running the systems, and what the published data does — and doesn't — say.
The data shows what works. Build the resume that benefits from it.