Job Seekers Try to Pass AI Screening With White Text Prompts

Key Takeaways

Last year, 82% of companies used AI to review resumes, per Resume Builder.
Candidates have been trying to game the system for years by hiding messages for AI chatbots in white text on their resumes.
This approach can backfire, experts say, as companies update their AI systems to detect it.

Companies are turning to AI recruiting software to help them sort through the mountains of resumes they receive for each job listing, which can be around 250 applications per opening. Last year, for example, 82% of companies employed AI to review resumes, according to a Resume Builder survey.

But as the hiring process becomes more automated, job seekers have been hiding white text in their resumes so that these AI programs can find them. In fact, according to The New York Times, hiding messages in white text in resumes, or “prompt hacking,” has become so common that companies are now updating their software to detect it.

ManpowerGroup, the biggest staffing firm in the U.S., told the NYT that it finds hidden text in around 100,000 resumes annually, about 10% of the resumes it scans with AI. Meanwhile, AI hiring platform Greenhouse, which reviews 300 million resumes every year for thousands of firms, disclosed to the NYT that 1% of the resumes it processed in the first half of the year contained white text messages.

And many recruiters hate it.

“It drives me nuts,” Farah Sharghi, a recruiter who has conducted over 10,000 interviews, told CNBC in September 2023 about prompt hacking. “There’s always a resurgence every year on TikTok where somebody makes a video about it, and it gets millions of views and people do it. It’s not going to work.”

Still, some job seekers say the method has resulted in success.

One anonymous recent college graduate told the NYT that she sent out 60 job applications in the psychology field this spring with her regular resume and received one interview. After discovering prompt hacking on social media, she applied to 30 jobs with her newly “hacked” resume and landed six interviews, ultimately securing a role as a behavioral technician at a medical company.

Another job hunter, a 50-year-old tech consultant, slipped some white text instructions into his resume too — and scored five interviews within a few days for tech-advising positions. However, this move wasn’t risk-free: One recruiter discovered the prompt hacking and turned him down because of it, he told the NYT.

How to make your resume AI-friendly without hidden text

According to Indeed, graphics can confuse AI scanning systems, so avoid complicating your resume with images unless you’re applying to a creative role, like a graphic designer, web designer, or animator.

Instead, keep things simple by using standard fonts like Times New Roman or Arial and making the headers of each section bold.

Target keywords are words and phrases that an AI system looks for when analyzing an application, per Indeed. These can be found in the job description and mentioned in context on a resume. Instead of general keywords like “team player,” emphasize specialized skills that the AI system is searching for, such as “data analysis” or “Python” for tech positions.

Employers should receive a version of the resume that is tailored to the job description. Staffers, a leading staffing firm, recommends that if you pull exact phrases from the job posting into your resume, aim for only a 60% to 85% keyword match, instead of copying the entire job description.

Resumes should be saved and submitted as a Word document or PDF to avoid confusing AI software. According to Jobscan, a PDF resume works the best because it preserves formatting.

Key Takeaways

Last year, 82% of companies used AI to review resumes, per Resume Builder.
Candidates have been trying to game the system for years by hiding messages for AI chatbots in white text on their resumes.
This approach can backfire, experts say, as companies update their AI systems to detect it.