Did AI Screen Your Job Application? The Rights Candidates Have Right Now

Say you apply for a job at 11 p.m. and the rejection lands before you’ve brushed your teeth. Probably nobody read that resume. Something likely scored it, and you have no idea what it looked for.

Most candidates assume there’s nothing to be done about that. In a few places, that assumption is now wrong. New York City, Illinois, Colorado, and California each have rules about employers using automated tools in hiring, and they give candidates different things: a notice, a way to ask for an accommodation, records that have to be kept, and, starting in January, a right to ask a person to look again. None of this is legal advice. It’s a map of what exists in late October 2026, so you know which questions are worth asking.

New York City makes employers tell you ten business days ahead

New York City’s Local Law 144 has been enforced since 2023, and it’s the most concrete of the four. If an employer uses an “automated employment decision tool” on candidates who live in the city, it needs an independent bias audit of that tool, and it has to post a summary of the results publicly.

The notice rules are the part you can use. According to the city’s own FAQ for employers, the employer must tell you it’s using the tool and which job qualifications or characteristics the tool will assess. The notice has to include instructions for requesting a reasonable accommodation under other laws, and it has to go out 10 business days before the tool is used. For job applicants, a notice on the employment section of the company’s website counts, and it doesn’t have to be tied to a specific role.

So before an application or a one-way video interview with a New York City employer, check the careers page footer for a link that mentions automated tools or a bias audit. If the tool is going to measure something your disability affects, the notice is where you find out how to ask for a different format.

Two caveats. The same FAQ says the law requires an audit but “does not require any specific actions based on the results,” so a bad audit doesn’t force an employer to stop. And enforcement has been thin. A New York State Comptroller audit released in December 2025 found that 75 percent of test calls to 311 about these tools were routed incorrectly, and that the city’s review of 32 companies’ bias audits turned up one compliance issue where the Comptroller’s office found at least 17. Complaints go through an online form linked from the city’s page on these tools, and you don’t need an account to file one.

Illinois made notice the law, and the rules for it stalled

Illinois amended its Human Rights Act with HB 3773, which took effect on January 1, 2026. It covers any employer with at least one employee in Illinois. Employers can’t use AI in recruiting, hiring, promotion, discipline, or discharge in a way that discriminates against protected classes, they can’t use ZIP codes as a stand-in for those classes, and they have to notify applicants and employees when they use AI to make employment decisions.

What the statute left out is how. The Illinois Department of Human Rights was supposed to write rules on when notice is required, how far in advance, and in what form. It formally proposed those rules on May 15, 2026, and then postponed the rulemaking on June 2, citing the need to keep working with other state agencies. As of late September, no new timeline had been published.

The practical result is a law with a clear duty and fuzzy mechanics. Some Illinois employers will post a notice; some will wait for the rules. There’s also federal pressure: a December 11, 2025 executive order told the Justice Department to challenge state laws that regulate AI, though one law firm pointed out that preempting state law usually takes an act of Congress. For a candidate, none of that changes the question you can ask, and in Illinois the recruiter should have an answer.

Colorado rewrote its law and moved it to January 2027

Colorado passed a broad state AI law in 2024 and then pushed back its start date. In May 2026 the legislature repealed and replaced it with SB 26-189, which the governor signed on May 14. Employment decisions are covered, and the obligations start on January 1, 2027.

This one is aimed at what happens after a decision, which makes it different from the other three. Under the new law, a company using a covered automated tool to materially influence a consequential decision has to give clear notice that it’s doing so. If the outcome goes against you, it has to send a plain-language description of the tool’s role within 30 days. You can ask to see the personal data it used, correct data that’s wrong, and request meaningful human review and reconsideration “to the extent commercially reasonable.”

That last phrase matters, and so does the enforcement model. Only the attorney general can enforce the law, and there’s no private right of action, so you can’t sue under it yourself. Still, a right to ask for a human second look is a real tool if an automated screen read your work history wrong. The attorney general is supposed to publish rules on the post-decision description by January 1, 2027, so expect the details to firm up around then.

California treats AI screening as a discrimination question

California went a different way. Rules from the state’s Civil Rights Council took effect on October 1, 2025, and they make clear that an automated decision system can violate state anti-discrimination law if it harms applicants based on protected characteristics. The examples in the rules are familiar from any modern application: screening resumes for particular terms or patterns, and analyzing facial expressions, word choice, or voice in online interviews.

Two parts help candidates directly. Employers now have to keep employment records, including automated decision system data, for at least four years instead of two, which matters if you ever bring a complaint and need the evidence to still exist. And the rules note that a tool measuring skill, dexterity, or reaction time may screen out people with certain disabilities, so employers may need to offer a reasonable accommodation. If a timed game or a reaction test is standing between you and a recruiter, asking for an alternative is reasonable in California.

What to ask a recruiter, and how to keep a record

You don’t need to know which law applies to ask good questions, and using AI responsibly across the hiring funnel starts with knowing where it’s being used on you. Most recruiters can answer these in a sentence, and the way they answer tells you something about the company. Keep the tone curious rather than adversarial:

1. Before the process starts: “Is any part of this process scored by an automated tool, like resume screening or video analysis? Is there a notice I can read?”
2. Before a one-way video or a timed assessment: “If this step is scored automatically, is there a different format available? I’d like to request an accommodation.” Only ask for an accommodation you actually need.
3. After a fast rejection from a Colorado employer, once the law is in effect: “Could you send the description of how the automated system was used in this decision? I’d also like to request human review.”
4. If something seems off: “Which parts of my application did a person review?”

Then write down what you heard. Note the date, the recruiter’s name, the question, and the answer, in the same place you keep notes for the rest of the search. A rights request six weeks later works better when you can say “your recruiter told me on October 30 that no automated scoring was used” than when you’re working from memory.

This is where a tool for your own notes helps. ExtraBrain is a desktop app for Mac and Windows that can transcribe a call live and generate a summary afterward, and saved sessions stay on your computer until you delete them. If you transcribe recruiter calls, get consent where the law or the other person requires it. Also check where your transcript goes, because cloud AI providers can receive transcript text when you use them. The privacy page explains which paths stay local. If you want a broader case for keeping these records, your interview transcript is career data, and it’s worth treating that way.

The next step is one email

If you’re applying to an employer in New York City, Illinois, Colorado, or California this week, look for the AI notice on their careers page before you apply. If it isn’t there, send the recruiter one line: “Before I go further, is any step in this process scored by an automated tool?” If you end up on a call about it, an ExtraBrain session summary gives you the dated answer for your notes. The question costs you nothing, and the answer tells you whether a person will ever see your application, and if one won’t, you can decide whether to spend your evening on it.