Can You Opt Out of AI Resume Screening? Honest Answer

Can you actually opt out of AI resume screening?

You can request an alternative human review in a handful of jurisdictions, but almost nowhere can you compel one today. The distinction that matters is between a duty to disclose and a duty to provide. New York City employers running an automated employment decision tool have to tell you the tool is in use and explain how to ask for an alternative selection process. The implementing rules then say, in plain terms, that nothing requires the employer to actually provide that alternative. Several published employer notices spell this out on the same page as the invitation to ask.

So the right most articles describe as an opt-out is really a right to be informed and a right to raise your hand. Whether anything happens next is a business decision, not a legal obligation. That changes on January 1, 2027 in California, where the state's automated decisionmaking rules give applicants a genuine opt-out from technology that replaces human judgment in hiring decisions, along with at least two ways to exercise it and a ban on retaliation for doing so.

One important carve-out cuts through all of this: if you need an accommodation because of a disability — a screen reader can't handle a timed game-based assessment, say, or a facial-analysis interview tool penalises a speech condition — you're no longer in the AI-policy lane. You're in disability law, which has real enforcement behind it and a long track record. That request is far more likely to be honoured than a general objection to being scored by software.

Which laws actually give you a right to request human review?

Five regimes matter for most applicants: New York City, California, Colorado, Illinois and the EU. Each grants something different, and only two of them put meaningful weight behind the words human review. New York City's law, in force since July 2023, requires notice at least ten business days before an automated tool is used on you, published on the careers page or sent by mail or email, with instructions for requesting an alternative. Enforcement sits with the city's consumer and worker protection department; penalties run to $500 for a first violation and up to $1,500 for each one after that. There's no route for you to sue personally.

California's rules are the strongest in the United States. They sit inside the state's consumer privacy law, which treats applicants and employees as consumers, and they reach for-profit employers that meet the law's thresholds — broadly, businesses over $25 million in annual revenue doing business in California. Before using the technology, those employers must explain how it works, what data feeds it, and what the alternative process looks like if you opt out.

Europe has had the sturdiest version of this right since 2018. Data protection law gives you the right not to be subject to a decision based solely on automated processing where it significantly affects you, plus the right to human intervention, to state your case, and to contest the outcome. The EU's top court has made clear that a token human clicking approve doesn't cure it — the reviewer needs genuine discretion. The AI Act's separate high-risk obligations for recruitment tools were pushed back to December 2027, but that delay changes nothing about the underlying data protection right.

Illinois is narrower than people assume. Its video interview law demands advance notice, an explanation, and your written consent before AI analyses a recorded interview, and deletion of the recordings within 30 days if you ask. It does not cover résumé screening, ranking tools, or automated assessments. A 2026 amendment to the state's human rights act added a notice duty for AI used in employment decisions — notice, not consent.

What actually happens after you ask to opt out?

Your application usually lands in a manual queue that nobody opens until the role is filled. That's the part the legal explainers skip. Recruiters coordinating a hundred applications a week work the ranked list the system hands them; an out-of-band email asking for special handling becomes a task with no deadline attached to it. Practitioners who advised employers when New York's law took effect reported that non-disability opt-out requests were vanishingly rare and raised awkward logistics questions nobody had answered. Rare requests don't get smooth processes built around them.

There's a second thing worth understanding: opting out doesn't pull your CV out of the applicant tracking system. Your record still gets parsed into fields, still gets indexed, and still gets pulled up when a recruiter runs a keyword search. What you're objecting to is the scoring or ranking layer sitting on top. If your CV parses badly or misses the language in the job ad, removing the score doesn't help you — a human searching that same database won't surface you either.

So the honest framing is this. Opting out doesn't buy you a careful human reading. It buys you a different, slower path to the same recruiter, minus whatever benefit of the doubt an automated match score might have given you. If your background is a clean fit on paper, the algorithm is more likely to be your friend than your obstacle.

When is opting out of AI screening actually worth it?

Opt out when the tool itself is the barrier, not when the concept bothers you. Three situations clear that bar. First, disability: if an assessment, a video interview analyser, or a timed game can't be completed fairly given your condition, request a reasonable accommodation and say so explicitly — that's an enforceable right in the US, the UK and across the EU. Second, a non-standard career shape that a scoring model reliably misreads: eight years abroad, a licence from another country, a career break, a title that means something different in your industry. Third, senior roles where the pool is small and a recruiter is reading everything anyway.

Don't opt out because you dislike the principle, because you read that the tools are biased, or because a role has hundreds of applicants and you're hoping for a shortcut. In high-volume hiring, the manual pile is the worst place to be. And frame the request as a request for review, never as a complaint about the employer's technology — the recruiter reading it has no authority over which tools their company buys, and you've just made yourself the difficult candidate before anyone has read your first bullet point.

My honest opinion, having watched how these mechanisms play out since New York's law landed: for most candidates the opt-out is paperwork theatre. Its real value is transparency. The notice tells you an automated tool is scoring you, what qualifications it claims to assess, and often which vendor's audit results are published. That's genuinely useful intelligence about what to put on the page — arguably more useful than the opt-out itself.

How do you word a request for human review?

Keep it short, name the law, and ask a concrete question the recruiter can answer in one line. Something like: 'Subject: Request for alternative selection process — Senior Analyst, req 4821. Hi Dana, I'm applying for the Senior Analyst role and I live in New York City. Under Local Law 144 I'd like to request an alternative selection process — a human review of my application without an automated employment decision tool scoring or ranking it. I'm glad to complete any equivalent non-automated assessment you use. Could you confirm you've received this and let me know how my application will be reviewed? Thanks, Sam.'

For a California employer after January 2027, swap the reference: 'Under the CCPA's automated decisionmaking rules, I'm opting out of automated decisionmaking technology for this application and requesting the alternative process described in your pre-use notice.' In the EU or UK, the phrasing is different again: 'I'm exercising my right under Article 22 not to be subject to a decision based solely on automated processing, and requesting human intervention in the assessment of my application.'

If a disability is the reason, say that plainly and skip the AI framing entirely: 'I'm requesting a reasonable accommodation for the assessment stage. The timed video assessment isn't accessible to me because of a disability. Could we arrange an alternative — a phone or live video interview with the hiring manager?' You don't have to disclose a diagnosis, and you shouldn't feel obliged to.

Send it to the recruiter listed on the posting, or to the careers or talent inbox if there's no name. Do it when you apply, not after a rejection — once the automated decision has been made, most of these mechanisms give you nothing to reverse. Colorado's law is the exception, since its human review path is triggered by an adverse decision rather than an up-front request.

Does asking to opt out get you quietly rejected?

No US or EU rule allows an employer to penalise you for asking, and California's rules explicitly prohibit retaliation — but proving it happened is close to impossible. There's no private right of action under New York City's law; only the city agency can enforce it. Colorado's version is enforceable by the state attorney general alone. Realistically, if a request gets your application shelved, you'll never know, and you'll have no practical remedy.

That asymmetry is exactly why the calculation matters. You're weighing a right with weak enforcement against a hiring process where a recruiter's convenience quietly decides a lot. In Europe the balance tilts further in your favour: regulators have taken the position that filtering applicants without genuine human review has been unlawful since 2018, and a coordinated enforcement push across national data protection authorities in 2026 put recruitment squarely in scope.

The pragmatic move is to make the request cost nothing. Polite, brief, no accusation, no legal threats, no citing of case law. A recruiter who reads three sentences and a clear question is far more likely to forward it than one who opens a paragraph about algorithmic bias. Save the formal language for a complaint to a regulator, if it ever comes to that.

What should you do instead of opting out?

Make the automated screen work for you rather than fighting it. The screening layer isn't reading between the lines — it's matching your text against a job description and checking whether your record parses into clean fields. Titles that match the posting's language, a skills section that uses the exact tool names in the ad, dates in a consistent format, and no critical information trapped in a header, a text box, or a two-column layout. Do that well and you don't need an exemption, because you're the candidate the ranking pushes upward.

Get an objective read before you apply, not after forty silent applications. Upload your CV for a free AI CV analysis and you'll see how it parses, which sections a recruiter's search would miss, and which keywords from your target roles simply aren't on the page. That's the same information the screening tool is working from — you just get to see it first and fix it. If you've already sent dozens of applications and heard nothing back, it's usually worth checking what's silently blocking your CV rather than assuming a machine has it in for you.

Then use the notice for what it's genuinely good for. When an employer discloses the qualifications and characteristics its tool assesses — and in New York City, they must — you're reading a public list of what the screen rewards. Map your CV to it. That's a better use of ten minutes than a request most employers are free to decline.

Frequently asked questions