The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Employers remain accountable when an AI tool screens, scores, ranks, or otherwise informs an employment decision. In the United States, civil-rights obligations do not disappear because a vendor supplied the software or a manager approved its recommendation. Human review can be a useful safeguard, but only when the reviewer can examine the evidence, question the output, consider accommodations, and explain the decision. A person’s presence in the workflow is not proof that the process is fair or lawful.
Who is accountable when AI makes a hiring decision?
The employer’s obligations follow the employment decision, not the identity of the software’s supplier. The EEOC says Title VII applies when automated systems make or inform selection decisions. Its materials discuss uses that include recruitment, hiring, monitoring, and firing. EEOC guidance on assessing adverse impact explains the federal employment-discrimination concern.
As an Amazon Associate I earn from qualifying purchases.
New York City’s Commission on Human Rights is explicit in its disability guidance: covered entities remain responsible for the actions and decision-making of AI systems and other technology they use, and cannot avoid liability for unlawful discrimination by saying the technology caused it. NYC Commission on Human Rights employment guidance addresses that responsibility. Vendor contracts may allocate duties between a company and supplier, but they do not by themselves remove an employer’s legal obligations to workers or candidates.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchAs EEOC Chair Charlotte A. Burrows put it in the agency’s October 28, 2021 announcement of its AI and Algorithmic Fairness Initiative: “While the technology may be evolving, anti-discrimination laws still apply.” This is an agency statement, not a court ruling, but it captures the core point: automating a step does not exempt the resulting decision from applicable law. EEOC announcement.
#1 Best Overall
Can an employer blame a hiring algorithm?
No. For the covered employer, outsourcing a score or recommendation does not erase responsibility for how it is used. That does not mean every adverse result from an algorithm is automatically unlawful; the question depends on the applicable law and the facts, including how the tool works and affects people.
Accountability is also distinct from technical causation. A vendor might have designed the model, but the employer chose to use it for a particular role and remains responsible for its selection process under applicable obligations. Employers need enough information to understand the tool’s intended use, limitations, inputs, and relevant evidence; a black-box result that no one can meaningfully assess is a governance problem, not a defense.
Does human review make an AI hiring decision fair?
Not on its own. The available official guidance does not establish that human managers are inherently fairer than AI systems, that AI is always more biased than people, or that adding a reviewer prevents liability. A reviewer who simply accepts a score without understanding it may add little protection.
Recommended Free Tools
A meaningful review should give the manager authority and practical ability to:
- Question a recommendation and examine the job-related evidence behind it.
- Consider relevant qualifications rather than treating a score or ranking as conclusive.
- Address accommodation needs and check whether the assessment itself created a barrier.
- Record the evidence considered and the reason for the final action.
- Correct an error or provide a route to challenge the decision where appropriate.
These are practical governance criteria, not a single legal test that applies identically in every jurisdiction. Human and automated processes should be assessed against the same central questions: are the criteria job-related, applied consistently, accessible, documented, and open to correction?
How AI-supported and manager-led processes differ
| Issue | AI-supported process | Human-manager process |
|---|---|---|
| Consistency | A tool can apply a stated process across many records, but consistency alone does not prove validity or fairness. | Judgment can vary by reviewer and context; the cited official sources do not quantify that variation. |
| Evidence | Scores and rankings need explanation, validation, and review for their effects. | Interviews, references, and impressions can inform decisions; employers should document job-related grounds and accommodations. |
| Bias and access | Data patterns or test and interface design can reproduce bias or disadvantage people with disabilities. | Human judgment can also produce discriminatory outcomes; a person making the decision is not automatically a safeguard. |
| Accountability | The employer remains subject to applicable obligations when a third-party tool supplies a score or recommendation. | The employer remains accountable for its own decision and process. |
| Challenge and correction | Provide required notice, accommodation or alternative-process routes, and practical ways to correct errors. | Identify the decision-maker and record the reasons and evidence considered. |
This is a practical comparison, not the result of a direct experiment comparing AI systems with managers.
What risks do employment decision tools create?
Discrimination in screening and selection
Automated systems can reproduce patterns present in data or introduce new barriers through their design. A tool that screens resumes, evaluates online presence, or analyzes video interviews may rely on signals that are difficult to interpret or may not measure the qualifications a job actually requires. The New York State Office of the State Comptroller’s 2025 audit describes these risks, including amplification of existing bias and weak transparency about tools’ capabilities and limits. New York State Comptroller audit of automated employment decision tools.
The EEOC’s account of its Title VII guidance says employers should assess whether automated selection procedures create disparate impact on protected groups. The “four-fifths rule” in the Uniform Guidelines is not a safe harbor: satisfying that measure does not guarantee that a selection procedure is free from prohibited disparate impact. Employers should not treat one statistical checkpoint as conclusive proof of fairness. EEOC 2023 Annual Performance Report.
Rank #3
Disability exclusion and inaccessible assessments
Automated tests and software can screen out a person with a disability who could do the job with or without reasonable accommodation. A system may also prompt disability-related inquiries. The EEOC and Department of Justice’s guidance highlights the need to consider accessibility and accommodation when using algorithmic tools. EEOC and DOJ guidance on the ADA and software algorithms.
Employers should ask whether an assessment measures a job-related ability or instead penalizes a disability-related difference in how someone communicates, moves, sees, hears, or completes a task. Candidates need a usable way to request accommodation before an automated assessment becomes a barrier.
What does NYC Local Law 144 require?
New York City’s Local Law 144 is narrower than a general rule for all workplace technology. It concerns covered automated employment decision tools used to screen a candidate or employee for an employment decision. Do not assume every software system used at work falls within that definition. The law applies to covered employers and employment agencies in New York City, not nationwide. NYC Administrative Code, Subchapter 25.
For covered use, the code sets requirements concerning bias audits and notice. Its online publisher cautions that the code may not reflect the latest legislation or rules, so consult the current city guidance and legal text for compliance details. The NYC Department of Consumer and Worker Protection AEDT page provides city information.
Audit and public summary
A covered tool must have a bias audit conducted no more than one year before use. Before use, the most recent audit summary and the distribution date of the audited tool version must be publicly available. An audit is a checkpoint required by the city rule, not a guarantee that a tool is valid, fair in every context, or free of legal risk.
Notice, accommodation, and information
Covered candidates or employees must receive notice at least ten business days before the tool is used. The notice must identify the tool’s use and the qualifications or characteristics it assesses, and allow the person to request an alternative selection process or accommodation. If the employer’s website does not provide the data type, source, and retention policy, those details must be available on written request within 30 days.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What the 2025 NYC audit says about enforcement
The New York State Comptroller’s 2025 audit reviewed the city Department of Consumer and Worker Protection’s oversight for July 2023 through June 2025. DCWP reported two AEDT complaints received during that period. The Comptroller found that the department had not investigated whether complaint intake worked effectively and that complaint-based enforcement could miss organizations that believed they were outside the law and therefore did not post notices or audits.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteIn a review of the same 32 company websites and audits, DCWP identified one potential compliance issue; the Comptroller identified at least 17 potential instances. Those are potential instances found in the Comptroller’s review, not adjudicated violations. The difference illustrates that formal audit and notice requirements do not guarantee effective detection or enforcement. The Comptroller’s audit sets out the findings.
How employers can govern AI employment tools
NIST’s AI Risk Management Framework offers voluntary guidance for incorporating trustworthiness considerations into the design, development, use, and evaluation of AI systems. It is not employment law and does not replace legal advice. NIST says the framework, released January 26, 2023, is being revised. NIST AI Risk Management Framework.
For an employer, a practical governance process can include:
- Define the decision. Identify the job decision the tool informs, the people affected, and the specific criteria the tool assesses.
- Understand the system. Obtain information about inputs, outputs, intended use, limitations, and the version in service; do not treat a vendor score as self-explanatory.
- Assess outcomes and access. Review whether the process disadvantages protected groups or people with disabilities, and provide an effective accommodation route.
- Assign real oversight. Ensure the reviewer can question the recommendation, consider other evidence, and change the outcome rather than rubber-stamp it.
- Document and correct. Preserve the basis for decisions, provide applicable notices, and establish a route for candidates or employees to seek correction or raise concerns.
- Revisit the process. Reassess when the tool, its version, role, or use changes, and check current legal requirements in each jurisdiction where it is used.
The legal discussion here is centered on U.S. federal guidance and New York City law. State, local, and international rules may add requirements. Confirm current NYC law and agency guidance before relying on notice or audit details, especially because the online code may not include the latest amendments.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
