AI is already screening you — parsing your resume, scoring your answers, ranking you against people you'll never see. Employers get playbooks, vendors get talking points, and the person being evaluated gets an email… maybe.
Last month I published a scorecard asking TA leaders nine questions about whether they can defend the AI they've adopted. This is the other side of the table — the same nine questions, asked by the person being screened. Because if an employer should be able to answer for their tools, you're allowed to ask about the ones pointed at you.
Reminder: The score isn't about you. It's about them.
Check a box for every answer you’re confident you can get. One point each.
How to use this: Not all nine get asked out loud in a first interview. Start with what's discoverable: the job posting, the application flow, the careers page, the privacy policy, the AI screening statement (if available). Save the direct questions (3, 4, 7) for the recruiter screen. Each is reasonable, professional, and revealing. How an employer reacts to being asked is data too.
[ ] 1. What does it evaluate and what was it trained on? Resume parsing, pre-screening questions, a video interview, a game-based assessment, a chatbot? You can't understand a score until you know what's being scored and what history the tool learned "good candidate" from.
[ ] 2. Has it been independently audited and can I see the results? Not the vendor grading its own homework. Right now, New York City is the only place that requires employers to publish bias audit results. Several states are moving the same direction but haven't gotten there yet. So, in most places, "can I see it?" is a voluntary question. That's the point: an employer with a clean audit has every reason to share it. (Want the deeper legal picture? Read: Algorithmic Bias Audit Compliance: Navigating the 2026 Frontier)
[ ] 3. If I'm rejected, can you explain how AI factored into it? In plain language, not a vague "the system made a determination." If nobody can explain the decision, nobody made one.
[ ] 4. Does a human make the final call? Does the tool rank and flag for a person, or can it reject you before any human reads a word? Those are different processes. You're allowed to know which one you're in.
[ ] 5. What are my rights where I live? Some jurisdictions require notice, consent, or disclosure before AI screens you — Illinois (as of this year, for AI in hiring decisions broadly), New York City's Local Law 144, California, the EU AI Act. Others have laws stalled, repealed, or still in rulemaking — this map redraws itself every few months. Your rights follow your location, not their headquarters. (Current state-by-state list: "A running list of states and localities that regulate AI in hiring" — HR Dive, continuously updated.)
[ ] 6. Who's accountable if the tool gets me wrong — you or your vendor? "That's the software" is not an acceptable answer. Someone chose it, configured it, and pointed it at your application.
[ ] 7. Was I told and can I contest it? Were you notified AI would evaluate you before it did? And if the machine's read of you is wrong, is there a path to correct the record — or just a voicemail box?
[ ] 8. Is anyone checking the outcomes? Does the employer monitor whether the tool screens fairly across groups over time or did they install it and walk away?
[ ] 9. Who owns this and who do I contact when something goes wrong? A name, a role, a process. If your application hits a glitch, an error, or an exposure, "we'll forward your message" is where accountability goes to die.
Scoring — how transparent is this employer?
7–9 answers: they can defend what they've adopted. 4–6: they've installed more than they can explain. 0–3: you just learned something about the process — and it wasn't in the job posting.
— Keirsten Greggs, Founder, TRAP Recruiter, LLC. 26 years in talent acquisition, building and transforming enterprise recruiting functions. Host, TRAP Chat. Creator of the TRAP Framework — Trust, Relationship Building, Accountability, Proactive Approach.