How to spot fake job applicants
A fake applicant is a person, or a stack of generated text, that will not survive a real conversation about the work. AI-written resumes and prompt injection make the paper look fluent. They do not make the work history real. You can catch most of this before you spend an interview.
What resume fraud looks like now
The old version was an inflated title. The new version is a resume written by a model to mirror your job description, sometimes with instructions hidden in the file that try to steer an automated screen. The prose is smooth. The projects are vague. Dates and employers do not connect.
Employers AI’s research on AI-generated resume manipulation covers the hidden-instruction case in detail. Use this page for the practical checks, and that paper for the attack itself.
Checks that take five minutes
Ask for one project they can explain without the resume in front of them. A real engineer can say what broke. A generated application repeats the job post.
Look for a detail you can verify: a product you can click, a talk, a commit, a customer. Fraud falls apart when it has to point at something outside the document.
- The resume matches your wording more closely than the person’s public profile does.
- Employment dates overlap in a way the person cannot explain, or every role is a few months with no manager you could name.
- The file contains hidden text, white-on-white copy, or unusual markup. That is a reason to review it by hand, not to auto-advance it.
Do not let the screen be the only reader
An automated screen is helpful when it scores against the job and shows its notes. It is a target when it trusts the resume as instructions. Keep a person able to open the original file. Reject hidden text rather than “fixing” it silently and continuing.
A short live or written screen that asks about their own system is the filter. Puzzle trivia is not. Fraudulent applicants are often fluent until the question requires a specific memory.
Common questions
- What is resume fraud?
- Resume fraud is an application that misstates experience, or that uses generated text and hidden instructions to pass a screen the person could not pass in conversation.
- How do I detect an AI-generated resume?
- Compare the document with a public profile, ask the person to explain one project from memory, and inspect the file for hidden text. Fluent writing alone is not proof either way.
- Should I reject everyone who used AI to edit a resume?
- No. People use editors. Reject applications you cannot verify, and any file that tries to instruct your screening system.