AI Video Interview Consent: A Candidate Checklist
Before recording an AI video interview, verify what the employer’s system does, what it analyzes, who receives the result, how long data is retained, and what alternative or accommodation is available. Illinois, Maryland, and New York City take different approaches: consent, facial-template waivers, and advance notice are not interchangeable.
Key takeaways
- “AI interview” is too broad to explain informed consent. Ask what the system collects, analyzes, produces, and shares.
- For Illinois positions, employers using AI to analyze applicant-submitted video must give advance notice, explain the AI’s operation and general evaluation characteristics, and obtain consent before evaluation. Illinois General Assembly
- Maryland’s specific rule concerns facial-recognition technology used to create a facial template during an employment interview; the waiver must be plain-language and contain specified details. Maryland General Assembly
- For covered tools in New York City, look for notice about the tool, assessed qualifications or characteristics, accommodation instructions, and bias-audit information; the framework is not generally framed as a candidate-consent requirement. NYC Department of Consumer and Worker Protection
- If recording, camera requirements, speech analysis, visual analysis, or timing could create a disability-related barrier, request an accommodation or alternative evaluation method before submitting.
- Save the invitation, consent language, employer explanations, vendor details, and every accommodation or deletion request in one dated file.
AI video interview consent means permission for an employer or technology provider to record and use a submitted video interview for stated purposes. Those purposes may include storing the recording, creating a transcript, analyzing answer content, evaluating voice or facial information, or helping decide whether an applicant advances. Meaningful consent requires enough information to understand what is collected, what the system evaluates, what output it creates, who receives it, and what happens if you decline. This is practical guidance, not legal advice. Our guide to AI voice interview preparation covers rehearsal, transcript checks, and privacy questions for conversational screens.
First, separate the decisions hidden in “AI interview”
“AI video interview” can describe very different operations. A platform may simply record answers for a recruiter, transcribe them, summarize responses, rank candidates, evaluate speech or facial movements, or create a facial template. The EEOC identifies recorded video interviews, facial expressions, voice, and movement as examples of AI use in hiring. That does not mean every platform performs all of those functions.
Before selecting “I agree,” separate the process into four questions:
- Is the employer recording and storing the video or audio?
- Is software analyzing answer content, a transcript, voice, facial information, movement, or another signal?
- What output does the system create: a transcript, summary, score, ranking, recommendation, or flag?
- Will that output influence the hiring decision, and does a human review the recording or automated result?
Then map the data flow. Ask whether the notice identifies the vendor, explains third-party access, describes what information is collected, and says whether the recording or derived data may be used for another purpose. You do not need the vendor’s source code to request a usable description of inputs, outputs, recipients, and retention.
Do not spend your preparation time imitating supposedly algorithm-friendly eye contact, facial expressions, or speaking patterns. First determine whether those signals are assessed at all. Prepare clear, job-related answers that a human reviewer can understand, and treat the consent notice as a process document rather than a performance puzzle. For recording logistics, use this video interview checklist for Zoom, Teams, and Meet.
Read the notice for seven concrete details
Before recording, look for plain-language answers to these questions:
- What is collected: video, audio, transcript, answer text, voice information, facial information, movement, or other signals?
- What does the AI evaluate, and what general characteristics or job-related criteria does it assess?
- Is the output a score, recommendation, ranking, flag, summary, or transcript?
- Can the output affect progression or screen out an application, or does it only organize information for a reviewer?
- Will a recruiter or another human review the submission and the automated result?
- Who receives the recording or derived data, including vendors, and how long is each item retained?
- How can you ask a privacy question, request an accommodation, or explore another assessment format?
Save the invitation, consent screen, privacy notice, instructions, employer contact details, deadline, vendor name, job posting, and any description of the AI criteria before submitting. Note whether the recording is scored, transcribed, used to train a model, or shared with a service provider if the materials address those points. A general website privacy policy may not answer interview-specific questions, so ask the recruiting contact to identify the relevant section.
Use a neutral message that asks for the missing facts:
“Before I complete the recording, could you confirm what parts of my video are analyzed, whether facial or voice features are used, how long the recording and derived data are retained, who receives the results, and whether I can complete an equivalent human-reviewed assessment instead?”
If the notice mentions particular functions, make the question more specific: “Does the system analyze only my transcript and answer content, or does it also evaluate voice, facial information, appearance, movement, or pauses?” Asking for clarification is different from alleging unlawful conduct. Compare the answer with the consent screen and keep both records.
Illinois: check notice, explanation, and consent
For Illinois-based positions, the Illinois Artificial Intelligence Video Interview Act requires an employer using AI to analyze an applicant-submitted video to notify the applicant before the interview, provide information about how the AI works and the general characteristics it evaluates, and obtain consent before the evaluation. An employer may not use the AI evaluation for an applicant who has not consented.
Read the notice for a meaningful general explanation. Does it distinguish answer-content analysis from analysis of vocal, facial, or other signals? The employer does not have to disclose a proprietary formula for you to ask what categories of data and characteristics are involved. If the notice says only that “AI may be used,” ask the employer to clarify the system’s function before recording and save the response.
Do not assume that consent to recording is identical to consent to AI evaluation. Ask whether the employer is requesting permission to store the recording, analyze it, or do both. Save the exact wording, including separate checkboxes, linked policies, and statements about categories analyzed or people who may receive the video.
The Illinois rule is a location-specific requirement concerning AI analysis of applicant-submitted video; it does not establish one nationwide consent process or guarantee an alternative non-AI interview route. The statute also places limits on sharing and allows applicants to request deletion of interview videos within 30 days. For a serious dispute, consider advice from a qualified attorney or an appropriate civil-rights agency.
Maryland: distinguish facial recognition from broader AI evaluation
Maryland’s specific employment-interview consent rule concerns facial-recognition technology used to create a facial template. The Maryland statute requires a plain-language waiver that states the applicant’s name, the interview date, consent to the use of facial recognition technology, and whether the applicant read the waiver.
This is narrower than a general question about whether an employer uses AI. A system that transcribes answers or evaluates answer content may raise different questions from one that creates a facial template. Read the waiver closely and save it with the invitation. If the employer uses broad language such as “automated analysis,” ask directly whether facial recognition is used and whether a facial template is created.
Do not treat a Maryland facial-recognition waiver as an explanation of every possible use of your recording. Ask separately about transcription, speech analysis, facial movements, retention, recipients, and hiring decisions. The goal is to identify each operation rather than assume one checkbox covers all of them.
New York City: look for notice and bias-audit information
New York City’s framework for covered automated employment decision tools is centered on notice and bias-audit requirements. The NYC Department of Consumer and Worker Protection says covered employers and employment agencies must complete a bias audit, make audit information publicly available, and provide required notices to candidates or employees.
When a New York City notice applies, check whether it identifies the tool, the job qualifications or characteristics it uses, and how to request an alternative selection process or accommodation. Covered employers generally must provide notice at least 10 business days before use. A bias-audit notice is not the same thing as a detailed explanation of every input or a promise that the system is accurate. Ask the recruiting contact which parts of the recorded interview are analyzed and how the result affects selection.
Keep the notice and any publicly available audit information with your application record. If the process refers you to a vendor page, save that page or its relevant text when you review it. Rules can change and may depend on whether the tool and employer are covered, so use the current city information rather than assuming every video platform falls within the framework.
Request an accommodation before recording
A one-way video format can create barriers unrelated to the job. You might have difficulty with speech, hearing, vision, motor control, facial movement, processing speed, captions, keyboard navigation, or the recording technology. The EEOC explains that AI tools can screen out qualified applicants with disabilities and that an alternative testing format may be needed.
Contact the person or team named in the invitation. Explain the barrier and the adjustment you need without volunteering unnecessary medical details. Focus on the assessment method rather than requiring yourself to disclose a diagnosis:
“I am interested in continuing with the application, but the automated video format creates an accessibility barrier for me. I need an alternative format or adjustment for this assessment because of a disability. What options are available, and who should handle the request?”
Depending on the situation, you could ask about captions, additional processing time, keyboard navigation, written responses, a live interview, or another effective adjustment. Do not assume one option must be available in every jurisdiction. Ask what process the employer uses, make the request early, keep the exchange, and ask how the accommodation will affect the assessment.
If the employer says the tool evaluates only answer content, ask how that is implemented. A transcript-based review raises different questions from a system that analyzes facial or vocal characteristics. A useful follow-up is: “Will the decision-maker see my transcript and qualifications, or will visual and vocal signals also be used?” Disclosure alone does not establish that an evaluation method is accurate, job-related, or free from discrimination. It also helps to read our guide to HireVue on-demand interviews, which covers answer frameworks, timed practice, technical checks, and retakes.
Treat accuracy and data-use claims cautiously
A notice may use reassuring language such as “objective,” “scientific,” or “accurate.” Those labels do not tell you which inputs are measured or whether the output is decisive. Ask the employer to translate the claim into a process description: What signal is collected? What category is assigned? Is the result used to reject, rank, or organize applications? Who can challenge an apparent error? A related read: our guide to whether you can use AI during a job interview, which explains why candidates should follow each interview’s AI rules and ask permission when unsure.
Be especially careful with vague claims about personality, emotion, engagement, or truthfulness inferred from a recording. Consent does not prove that an AI system is accurate, bias-free, or legally compliant. Federal employment protections still apply when employers use AI, as the EEOC explains in its worker guidance. If the response remains general, record that limitation rather than trying to reverse-engineer the system through your performance.