3 Tactics Candidates Use to Outsmart Your Screening Tools

Code that makes no sense. Claims and evidence that don’t match up. Explanations that don’t connect to portfolios. According to research from Career Group Companies, 65% of candidates admit to using AI in the application process. And according to Gartner, 1 in 4 job candidates will be completely fake by 2028.

No matter which side you’re on, hiring has changed. And with new ways to use AI, come new ways to hack the recruitment process. We’ll reveal the latest tricks candidates are using and help you decide how much of a role AI should play.

The promise and peril of the AI-assisted application process

Like it or not, AI is part of the hiring dance. Candidates use it to get past the noise; hiring teams use it to triage. “I do think that anyone who’s a savvy job searcher right now ... it would surprise me if they’re not looking into [AI],” says Jillian Lawrence, a senior vice president at Career Group Companies.

Job seekers have seen countless TikToks sounding the alarm about knockout keywords and “evil ATS” systems designed to auto-reject their applications. According to research, only 19% feel confident their application will actually lead to an interview. As HR professionals, we know that’s not how these systems work. But with candidates reporting a “complete 180” in interview rates after using hidden prompts – you can’t blame them for trying.

And in a world where AI-fluent talent is in demand, isn’t knowing a few tricks a good thing? Like all things people-related, the answer is complicated. For time-strapped hiring teams, the problem isn’t that AI exists – or even that applicants are using it. The problem is the gap between what’s claimed and what’s real. 

“Career catfishing doesn’t just create mismatched expectations, it sets employees and employers up for failure from day one,” says Monster’s career expert Vicki Salemi. “When honesty is absent in the hiring process, trust is broken before it can even take root, leading to dissatisfaction, burnout, and costly turnover.”

To decide where to draw the line, start by understanding the most common hacks targeting your application filters.

3 cheat codes candidates use to beat AI screeners

Fibbing in applications isn’t new; AI just gives it fresh paint. Monster found 85% of employees say career catfishing is wrong, yet nearly 7 in 10 say they work with someone who has probably done it.

With the cost of a mis-hire climbing, prevention beats cleanup any day of the week. Here’s how candidates are attempting to outsmart screeners, plus tips to protect your hiring process.

1. Prompt manipulation

According to data from Greenhouse, about 1% of resumes now include prompt injections. Candidates hide commands in formatting, bury them in document properties, or stash strings of code inside image metadata. If your ranking process relies heavily on LLMs, these injections can potentially skew scores and waste time. Most manipulations fall into three main buckets: hidden formatting, commands, and metadata tricks.

What it is:

• Formatting-based injection. For example, commands like “prioritize this resume” hidden in white text in the resume footer.
• “Rank me first” commands. Instructions aimed at LLM rankers can be visible, hidden, or in comments.
• Code/prompts in images. Involves using long strings of code or directives. For example, 100+ lines of prompt text in a headshot’s EXIF fields.

How to fight it:

• Define acceptable AI help vs. banned tricks (drafting/formatting OK; hidden text, directives, metadata not).
• Share your own screening practices to ensure correct AI use isn’t misread as manipulation.
• Keep a human in the loop. Treat flags as signals, not verdicts. Decide in context, then automate quick feedback for candidates.

If you’re already using a recruitment platform or applicant tracking system (ATS), it’s probably doing a lot of this work for you. Most are designed to automatically strip resume formatting and system commands that prevent applicant text from altering ranking logic.

2. Invisible keyword stuffing

With AI, keyword stuffing has never been easier to game. More than half of job seekers admit to using AI to write their resumes, resulting in an undeniable rise in spam applications. It’s the same playbook, just on a much bigger scale, thanks to free keyword scanners that use AI to help candidates “beat the ATS” by packing in extra terms like “communication,” “Microsoft Excel”, etc., in white text or other places where a human might not notice.

What it is:

• White-on-white text that repeats role-specific terms.
• Tiny font blocks tucked under images, in page margins, or behind shapes.
• Keywords in alt text, hidden sections, or document properties that don’t show properly on-screen.

How to fight it:

• Warn applicants about fraud detection tools. Help AI-savvy candidates understand where the line is so they can use their skills where they count.
• Cap the weight of raw keyword matches. Use a screening process that prioritizes skills, evidence, and relevance over broad term counts.
• Use detections as signals, not verdicts. Pair flags with human review and relevant candidate feedback.

The good news is, most ATS systems have countermeasures in place to prevent this from becoming an issue. Look for tools with features designed to catch invisible text, density spikes, or suspiciously perfect keyword matches. By using AI to flag the funny stuff, you can spend more time on applicants that show real effort.

3. Catfishing AI videos

Video is a great signal, until it isn’t. With more companies leaning on AI interviews, candidates are testing the limits with filters, voice changers, and even deepfake avatars. In some nightmare cases, these fake candidates are not only taking the interview, they’re also landing the offer.

What it is:

• Deepfake and liveness tricks using face filters, voice changers, or avatars to alter sound and appearance during one-way and/or AI interviews.
• Scripted responses with teleprompters that hide eye movement and smooth over pacing and tone.
• Over-edited takes with subtle jumps, perfect lighting/audio, and metadata that doesn’t match the recording platform.
• Identity mismatches where the voice, timestamps, or background don’t line up with the candidate’s profile information.

How to fight it:

• Delay video screening. Use questionnaires and phone pre-screens early and save video for later stages, like interviewing.
• Publish acceptable AI use. Allow practice and formatting help, but don’t permit face/voice alteration, scripted teleprompters, or off‑screen answer generators.
• Add light liveness checks. Record in-platform and ask for a simple gesture, like naming today’s date.

Remember, no matter how much AI makes its way into the HR process, candidates still prefer human conversation. Research shows that when candidates see “AI interview” in a job ad, application intent drops by about 7%. Other surveys have found that 62% of job seekers say they’re more likely to apply when in-person interviews are required. Whenever possible, go with a human for high-impact touchpoints like interviews.

Practical tips to catch the hacks and keep the humans

Expect candidates to use tools just as you do. With different views on what’s “fair,” anchor your process in human-in-the-loop checkpoints where judgment matters most.

• Set the tone: Use AI to triage, not auto‑nuke. Document decisions, share light feedback, and move strong signals to a human fast.
• Detection and filtering: Auto‑flag hidden text, invisible keywords, and anomalous metadata; quarantine files for human review.
• Streamline your pipeline: Almost 1 in 4 recruiters/hiring managers struggle to track or manage candidates. Instead of blindly outsourcing to AI, save time by using one system for all things hiring.
• Transparent criteria and human oversight: Publish what’s scored and how; ensure humans make final decisions.
• Clarify policy upfront: State where AI use is acceptable (e.g., drafting, brainstorming) and where independent reasoning is required.
• Delay face/video steps: Use audio or phone pre‑screens to limit early visual bias; reserve video for later stages.
• Two‑way touchpoints and feedback: Convert one‑way steps into application add‑ons, then follow with live Q&A; offer brief score/feedback to keep the experience positive for candidates.
• Measure onboarding outcomes: Less than 30% of teams use time‑to‑productivity in the first 90 days. Pair feedback with milestone tracking so you can connect screening quality to ramp time.‍
• Close the tooling gaps: 28% of hiring teams say their systems don’t integrate. Connect ATS data to onboarding to spot patterns in candidate quality and retention.

The AI hiring checklist: do this, skip that

The win isn’t “AI everywhere”; it’s AI in the right moments. Here’s a scannable checklist to help you find the right guardrails.

Use AI for the right reasons:

• Prep better interviews: AI summaries, transcripts, and past touchpoints can help human interviewers arrive sharp, not “winging it.”
• Score what matters: Ditch the keyword bingo. Rate skills, relevance, and overall fit with clear weightings behind every score.
• Quick pre-screens, not unnecessary hurdles: Short, role-specific questionnaires beat meandering chatbot dialogues. Use automated ranking to surface the best without auto-rejecting sleeper talent.
• Protect integrity without “gotchas”: Use AI to flag copy-pasted/keyword stuffed/AI-generated content so you can review suspicious patterns, without penalizing legit applicants.
• Anchor to clear requirements: 79% of employers already rely on clear job descriptions to bring in qualified candidates – use those as the backbone for automated screening criteria.

Don’t use AI for the wrong jobs:

• Don’t replace conversation: Talk to people. An AI avatar can share information and answer basic questions, but it can’t create trust or assess critical thinking.
• Don’t test with Google-able prompts: If the internet can answer it, you’re assessing copy/paste, not competence. ChatGPT-proof your interview questions by tying them to your tech stack and role requirements.
• Never auto-nuke applications: Treat flags as clues. Always use humans to review, document, and make the final decision.
• Don’t ignore enablement: 26% of teams lack training/resources for new hires. Use AI and automation to support onboarding, not replace it. Track whether AI assistance actually improves time‑to‑productivity.

Make AI hiring faster and fairer with the right guardrails

AI can make your hiring faster and fairer – or fuzzier and riskier – depending on how you use it. Publish clear AI guardrails, weight real evidence over surface language, delay video until it adds value, and always keep a human in the loop. 

That’s how you catch the hacks without punishing honest, AI‑literate candidates. 🫶

And Breezy is here to help. As soon as a candidate applies, Breezy’s Resume Audit checks for AI‑generated or copied content, bot‑like timing, and suspiciously perfect JD matches. It then surfaces an originality rating at the top of the résumé, plus details in the profile panel so human reviewers can make better decisions.

AI applications shouldn’t slow your team down. Try Breezy Intelligence today.