Policy · United States

AI hiring and fair employment

Written first for candidates who were interviewed by our AI and want to know what it did. The last two sections are for employers running the platform.

Last updated August 2026

The short version

An AI called you, asked the questions your employer wrote, and scored your answers against a rubric your employer defined. It recorded the call, and it told you so before it started.

The interview itself decided nothing. No interview score rejects a candidate; a person at the company reviews every scored candidate. One thing does run automatically though, and you should know it: if your résumé scored below the threshold that employer set, the application can be closed before anyone calls you. You can ask for a human interview, and you can ask what we hold about you.

This summary is here to be read. It does not replace the full text below, which is what governs.

What happened on your call

You applied for a role. The employer set a threshold, you met it, and the system placed an outbound call inside the calling window that employer configured. The voice on the line was software, and it said so.

The call was recorded and transcribed. Before the interview started, the AI told you the call was being recorded and asked you to confirm. We do that on every call in every state rather than only in the states that require it, because varying the script by area code is a worse idea than simply always asking.

Afterwards, the transcript was scored against the employer's rubric, and the score plus the reasoning behind it was attached to your application. You can ask for a copy of both.

What the AI scores, and what it ignores

The AI scores what you said against the competencies the employer defined for the role. Each score carries the part of your answer that produced it, so the reasoning can be read rather than guessed at.

It does not score, and is not given:

Your accent, speech rate, fluency, or how "confident" you sounded.
Your face or any image of you. The interview is voice only. No facial analysis runs at any point, in any state.
Your race, color, religion, sex, sexual orientation, gender identity, national origin, age, disability, veteran status, or genetic information.
Any voluntary self-identification you provided. That data is stored separately from the hiring record, is used only for aggregate reporting, and is never shown to the AI or to the people reviewing you.

Scoring runs deterministically, at temperature zero. The same transcript produces the same score every time it is scored. That is a prerequisite for auditing a tool for bias at all, because a tool that answers differently on Tuesday cannot be meaningfully tested on Monday.

Where automation stops, and where it does not

There are two screens, and they behave differently. Being straight about the difference matters more than a reassuring summary would.

The interview never rejects you

After the AI interview, the score and a recommendation are shown to a recruiter. The recommendation is not executed. No interview score moves a candidate to rejected; scored candidates are routed to a person for review. Advancing, holding and rejecting after an interview are all actions a human takes, and every stage change is written to an audit trail with the name of the person who made it.

The résumé screen can

Before any call, your résumé is parsed and scored against the requirements the employer set for the role. If that fit score falls below the threshold the employer configured, the system can close the application automatically, and no human will have read it. That is an automated rejection and we are not going to call it anything else.

Three limits are built around it:

It only ever applies to an application still sitting at the first stage. Once a recruiter has made any decision, a later rescore cannot overturn it.
If your résumé could not be read cleanly, or the parser had low confidence in what it extracted, the automatic rejection is withheld and the application is parked for a human instead. A file the machine struggled with is not treated as a weak candidate.
The threshold is the employer's number, not ours, and they can set it to zero so that nothing is ever closed automatically.

If you believe your résumé was misread, say so through the contact page. You can ask for the parsed version we generated, correct it, and ask the employer to reconsider.

Asking for a different format

You can ask to be interviewed by a person instead of the AI, and you do not have to explain why or disclose a disability to ask. The request link is on your application status page, or you can reply to any message you received about the role.

If a phone interview is not workable for you for any reason, say so. Under the ADA the employer is responsible for providing reasonable accommodation, and the platform is built so that switching a candidate to a human interview does not disadvantage them in the pipeline: the role reverts to the employer's standard human process and the candidate is scored on the same rubric.

What you can ask for, by where you are

US hiring law is not one rule, it is a patchwork, and the state you were in when we called you decides which of these applies to you. Find your row. If your state is not listed, the two rows marked Anywhere in the US still cover you.

New York City
Local Law 144 (AEDT)
An independent bias audit of the automated tool within the last year, published results, and at least 10 business days notice before it is used on you.
Us + employer
Illinois
AI Video Interview Act, 820 ILCS 42
Notice that AI will evaluate the interview, an explanation of what it measures, your consent before it runs, and deletion within 30 days of your request.
Us + employer
Colorado
SB 24-205 (AI Act)
Notice that a high-risk AI system is involved in the employment decision, an explanation if it goes against you, and a route to correct the data behind it.
Us + employer
Maryland
HB 1202
Your written consent before any facial recognition runs during an interview. TalentValue365 interviews are voice only and never use facial recognition.
Us
California, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania, Washington
All-party consent recording statutes
Every party has to agree before the call is recorded. The AI states that the call is recorded and asks you to confirm before the interview begins, on every call, in every state.
Us
California
CCPA / CPRA
Know what was collected about you, get a copy, correct it, delete it, and limit how sensitive information is used. Applicants count as consumers.
Us + employer
Anywhere in the US
Title VII, ADA, ADEA (EEOC)
Selection that does not discriminate on a protected basis, and a reasonable accommodation for the interview format, including a human interviewer instead of the AI.
Us + employer
Anywhere in the US
TCPA
The interview call is placed only after you apply and supply your number for this purpose, inside the calling window the employer configures.
Us

Candidate rights by jurisdiction, the law that grants them, and who is responsible for delivering them
Where you are	What applies	What you can ask for	Who owes it
New York City	Local Law 144 (AEDT)	An independent bias audit of the automated tool within the last year, published results, and at least 10 business days notice before it is used on you.	Us + employer
Illinois	AI Video Interview Act, 820 ILCS 42	Notice that AI will evaluate the interview, an explanation of what it measures, your consent before it runs, and deletion within 30 days of your request.	Us + employer
Colorado	SB 24-205 (AI Act)	Notice that a high-risk AI system is involved in the employment decision, an explanation if it goes against you, and a route to correct the data behind it.	Us + employer
Maryland	HB 1202	Your written consent before any facial recognition runs during an interview. TalentValue365 interviews are voice only and never use facial recognition.	Us
California, Florida, Illinois, Maryland, Massachusetts, Montana, New Hampshire, Pennsylvania, Washington	All-party consent recording statutes	Every party has to agree before the call is recorded. The AI states that the call is recorded and asks you to confirm before the interview begins, on every call, in every state.	Us
California	CCPA / CPRA	Know what was collected about you, get a copy, correct it, delete it, and limit how sensitive information is used. Applicants count as consumers.	Us + employer
Anywhere in the US	Title VII, ADA, ADEA (EEOC)	Selection that does not discriminate on a protected basis, and a reasonable accommodation for the interview format, including a human interviewer instead of the AI.	Us + employer
Anywhere in the US	TCPA	The interview call is placed only after you apply and supply your number for this purpose, inside the calling window the employer configures.	Us

This table describes what the named laws require. It is not legal advice, and it is not a claim that any particular employer using TalentValue365 has met its own obligations. To exercise any of these, contact the employer you applied to, or us, and we will route it.

Adverse impact monitoring

The platform computes selection rates by stage from stage-transition history and applies the four-fifths rule from the Uniform Guidelines on Employee Selection Procedures, so an employer can see whether one group is advancing at less than 80% of the rate of the highest group.

Two honest limits on that. First, it can only measure groups for which candidates voluntarily self-identified, so thin response rates make it less reliable. Second, this is monitoring, not a bias audit: it tells an employer where to look, and it is not a substitute for the independent audit that NYC Local Law 144 requires.

Current independent audit status: auditor and audit date to be published here. Until that is filled in, employers hiring in New York City should treat the Local Law 144 audit as outstanding and not rely on this page for it.

If you are the employer

The platform gives you the tooling. Several obligations remain yours, and the distinction matters because candidates will ask you, not us:

Serving the required notices to candidates, including the NYC Local Law 144 10-business-day notice and the Illinois consent and explanation.
Commissioning the independent bias audit if you hire in a jurisdiction that requires one, and publishing its results.
Deciding the rubric and the thresholds. The AI measures what you tell it to measure, so a rubric containing a proxy for a protected characteristic will produce exactly the problem you are trying to avoid.
Providing reasonable accommodation, and making the route to ask for it visible.
Keeping records for the periods your jurisdiction requires.

Reaching a human

To ask what the AI recorded about you, request a human interview, or raise a concern about how you were assessed, use the contact page. Tell us the company you applied to so we can route it. If your request is about a decision, the employer is the one who made it and we will put you in front of them.

You can also contact the US Equal Employment Opportunity Commission directly at eeoc.gov. Nothing on this page limits that right.