AI in Hiring: How Employers Screen You and What Rights You Have

If you apply for jobs online in the United States, there is a good chance software reads your application before a person does. That software can range from a simple keyword filter to systems that score recorded video answers or game-based assessments. Knowing how these tools work, and what the law says about them, helps you present yourself well and recognize when something has gone wrong.

The main types of hiring AI

Applicant tracking systems

Most medium and large employers use an applicant tracking system (ATS) to collect and organize applications. Many simply store résumés and let recruiters search them. Increasingly, ATS platforms add features that parse résumés into structured fields, rank candidates against the job description, or flag "best matches." The ranking is only as good as the criteria behind it, which may weigh keywords, job titles or years of experience.

Chatbot recruiters

Some employers, especially in retail, hospitality and logistics, use conversational bots to answer questions, ask screening questions (availability, work authorization, certifications) and schedule interviews. These bots often apply knockout rules: answer "no" to a required question and the application may end there.

One-way video interviews

You record answers to preset questions with no one on the other end. Some platforms only pass recordings to human reviewers; others transcribe and analyze the content of answers. A few vendors once marketed facial-expression analysis, a practice that drew criticism from researchers and regulators and has been scaled back by some providers.

Online assessments

Skills tests, situational judgment questions, personality inventories and puzzle-style games can generate scores that feed into ranking. Some are scored by fixed rules, others by models trained on data about past employees.

Why fairness is a real concern

Automated tools can reproduce bias in their training data. A widely reported example: Reuters reported in 2018 that Amazon had scrapped an experimental recruiting tool after finding it penalized résumés that included the word "women's," because it learned from a decade of mostly male applicants. The tool was never the company's sole hiring method, but the case became a standard illustration of the risk.

Other concerns include:

• Proxy discrimination: factors like zip code, gaps in employment or certain schools can correlate with race, age, disability or caregiving status.
• Accessibility: timed games, video analysis and speech transcription may disadvantage people with disabilities or non-native accents.
• Opacity: candidates rarely learn why they were screened out, which makes errors hard to catch.

What US law says

There is no single federal AI hiring law, but existing laws apply fully.

The key principle is that an employer cannot escape responsibility by pointing to a vendor's algorithm. If a tool screens out a protected group at a disproportionate rate and the practice is not justified as job-related, the employer may be liable.

New York City's Local Law 144, which the city began enforcing in 2023, requires employers using covered automated tools for hiring or promotion decisions to have an independent bias audit within the prior year, publish a summary of the results, and notify candidates in advance. The city's AEDT page explains the details.

Other states have passed or proposed broader AI laws covering "high-risk" decisions such as employment, and some timelines have shifted after passage. Because this area changes frequently, check your state attorney general's or civil rights agency's website for current rules.

What you can do as a candidate

Before you apply

• Write for both software and people. Use a clean format with standard headings ("Experience," "Education," "Skills"), avoid putting key information in images or text boxes, and mirror the posting's terminology where it honestly applies.
• Read the application carefully. Knockout questions often decide your fate. Answer accurately; if a requirement is "preferred" rather than "required," say what you do have.

During assessments and video interviews

• Ask what will be evaluated, especially for recorded interviews. Some employers explain this upfront, and in some places they must.
• Request accommodations if a format creates a barrier because of a disability: extra time, an alternative format or a live interview. You do not have to disclose a diagnosis in detail; describe the barrier and the adjustment you need.
• Control what you can: good lighting, a quiet room, a stable connection, and concise, structured answers.

After a rejection

Rejections are routine and usually not evidence of anything. But if you see a pattern, such as being screened out instantly by an employer whose requirements you clearly meet, or a tool that seemed impossible to complete with your disability, you have options:

1. Ask the employer how applications are reviewed and whether a human can look at yours.
2. Keep records: the posting, your application, any assessment emails and dates.
3. Contact the EEOC or your state or local fair employment agency. Federal charges must generally be filed within 180 days of the act, extended to 300 days in many states with their own agencies; the EEOC explains the process.

Questions worth asking employers

Recruiters often appreciate thoughtful process questions. You might ask:

• "How are applications reviewed at the first stage: by a person, by software, or both?"
• "Is there an alternative format for the assessment if I need an accommodation?"
• "Who reviews recorded interviews?"

Employers that use AI responsibly can usually answer these clearly. Their answers tell you something about the organization, too.

AI screening is now part of the hiring landscape, but it does not change the basics: a clear, honest application, a well-prepared interview and the right to be judged without discrimination. Knowing how the tools work just makes it easier to get your application in front of a person.