




AI-analyzed video interviews are becoming part of hiring in New Jersey. Employers use software to assess facial expressions, speech patterns, and recorded answers. Proposed bill A3911 focuses on transparency.
When employers use AI in video interviews, disclosure rules shape how well applicants understand the process and whether it meets legal standards.
Applicants often complete recordings without clear notice that automated systems will assess their responses. In the cases we handle at Brandon J. Broderick, that lack of disclosure appears even when employers treat these tools as routine screening. A3911 addresses this by requiring notice of AI use. It changes how hiring processes are structured and documented. A standard format takes on legal significance when algorithmic bias is involved.
In this guide, we explain what A3911 proposes, how AI-analyzed video interviews are regulated, what employers must disclose, and when to speak with an employment lawyer in New Jersey.
Hiring has shifted toward recorded video interviews. Applicants record answers on their own time, often through a platform that later scores the responses. Some employers rely on artificial intelligence to review those recordings, with systems assessing word choice, tone, pace, facial movement, or other observable traits, sometimes before any human review takes place.
Roughly 83% of employers and up to 99% of Fortune 500 companies use automated tools to screen candidates. One in four medium-sized employers uses automation or AI as part of the hiring process.
A human reviewer may never see the first round of candidates.
New Jersey Assembly Bill A3911 targets this setup. It applies when an employer asks an applicant to submit a recorded video interview and uses artificial intelligence to evaluate it. Both pieces must be present. A live call without AI review falls outside the core of the bill. A resume scanner or keyword tool without a video component isn’t the focus.
A3911 would amend Title 34 of the New Jersey statutes, which governs labor and employment. It defines “employer” broadly. Coverage extends to corporations, limited liability companies, nonprofits, educational institutions, and other entities that employ workers in the state.
A resume shows education and work history. But a video shows voice, cadence, facial expressions, eye movement, lighting, and background. It can reveal disability-related traits or age-related cues. An algorithm built to score culture fit or “engagement” might weigh those signals. An applicant doesn’t know which signals are being measured or how they affect the score.
A3911 treats that lack of visibility as a legal problem. It does not ban AI in hiring. It focuses on disclosure, consent, and limits on how employers use and store recordings. The bill also adds a reporting requirement tied to race and ethnicity when AI determines who moves forward in the hiring process.
Courts already treat hiring tools as part of the selection process. Under Title VII of the Civil Rights Act of 1964, any tool used to screen applicants must avoid unlawful discrimination. If a selection tool screens out minority groups or shows evidence of racial bias, the employer must address that impact and show the tool is based on legitimate criteria. The rule applies to written tests and algorithms alike.
New Jersey law reaches the same result through the New Jersey Law Against Discrimination. The NJLAD bars discrimination in hiring based on race, national origin, sex, disability, age, and other protected traits. In 2025, the New Jersey Division on Civil Rights issued guidance on algorithmic decision-making. Automated systems remain covered by the statute.
“The decision to speak up is powerful. But knowing what happens after — and how to protect yourself — is just as critical.”
— Olivia Rhye
A3911 requires employers to speak plainly before asking for a recorded video. General language buried in an application isn’t enough. Notice must come before the interview takes place.
The bill requires:
Employers must explain how the AI works at a general level. This doesn’t require disclosing proprietary code. An applicant should understand how the system scores answers, measures speech patterns, tracks facial movement, or flags keywords. That explanation should also clarify whether visual factors are considered, such as religious clothing. In the cases we handle at Brandon J. Broderick, this type of clarity often becomes a key issue when hiring practices are reviewed.
Clarity matters because the interview isn’t interactive. In a live setting, an applicant can ask questions or adjust responses based on feedback. A recorded session offers no back-and-forth. If an algorithm scores the recording, the applicant never sees how the score is formed.
A3911 bars employers from using AI to evaluate an applicant who has not agreed to it. This creates a choice. An applicant can decline and still expect the employer to handle the application without AI analysis, or not proceed with the video process at all.
Federal law already places limits on how selection tools operate. The EEOC’s Uniform Guidelines on Employee Selection Procedures require employers to validate selection tools that have a disparate impact.
If a device screens out candidates in a way that correlates with a protected trait, the employer must show the device is consistent with business necessity. A3911 doesn’t change this burden. It adds a front-end disclosure rule so applicants know what is happening before their data is analyzed.
A simple example shows the difference. An employer asks candidates to record answers to three questions. The platform analyzes tone, eye contact, and word choice, then ranks candidates. Under A3911, the employer must tell the applicant about the AI analysis, explain what types of traits are being evaluated, and obtain consent before the recording. Without those steps, the employer cannot use the AI output to evaluate the candidate.


Recorded interviews create a record that is easy to store and share. A3911 places limits on both. It treats the video as sensitive information tied to a job application.
An employer may share the recording only with a service provider whose expertise or technology is necessary to evaluate the applicant. In our experience, many employers rely on outside vendors to host or process these interviews. A3911 allows limited sharing for that purpose. It prohibits broad distribution beyond what is needed to evaluate the candidate.
The bill also creates a clear deletion right. If an applicant asks for deletion, the employer must act. The obligation extends beyond the employer’s own systems. Key requirements include:
Those steps turn retention into a defined duty. A recording does not sit on a server indefinitely. Once an applicant requests deletion, the clock starts.
This approach aligns with other state efforts. Illinois adopted the Artificial Intelligence Video Interview Act, which also requires notice, consent, limits on sharing, and deletion upon request. Recorded interviews carry more personal detail than a written application.
Employers already handle sensitive applicant data under strict rules. Social Security numbers, background checks, and medical information are subject to clear requirements, including data breach notifications if that information is exposed. Videos add another category, and A3911 places it within a defined set of obligations.
Vendors also play a role. If a third-party platform hosts or analyzes the video, it must follow the employer’s deletion instructions. This helps ensure no duplicate copy remains. The bill ties responsibility to the full chain of custody.
A mishandled video record could intersect with discrimination claims under the NJLAD or with other privacy-related claims, depending on the facts.
A3911 goes beyond notice and privacy. It requires employers to track outcomes tied to AI-analyzed video interviews and report them to the state. The focus is on race and ethnicity at key decision points.
An employer must track and report the race and ethnicity of applicants if they are:
Those figures are reported annually to the New Jersey Department of Labor and Workforce Development. The department reviews the data and submits a yearly report to the governor and the Legislature.
Existing law already addresses discriminatory outcomes. The NJLAD prohibits discrimination in hiring based on protected traits. A plaintiff can bring a claim if a hiring practice excludes candidates in a way tied to race, national origin, sex, disability, or other protected categories. Federal law under Title VII follows the same path.
AI tools fall within those rules. The EEOC has made that point in its guidance on artificial intelligence. An automated system that screens applicants is still a selection procedure. If it disproportionately excludes a protected group, the employer must validate or adjust the tool.
A3911 supports that process by requiring data collection at the front end. Without it, patterns remain hidden. Applicants may never see the reports. The state will. Over time, those reports build a record of how AI tools perform across industries and job categories.
If you have questions about how these rules apply to your situation, it may be worth taking a closer look. Contact us today for a free consultation.

Stop wondering about your rights or if you'll be taken seriously. We treat every client with respect, urgency, and honesty. Our lawyers will listen, explain your legal options, and fight for the outcome you deserve.