Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Employers can audit AI hiring tools for bias by mapping every decision the software influences, measuring outcomes for relevant groups at each consequential stage, checking accuracy and accessibility, and documenting corrective action. In New York City, covered automated employment decision tools (AEDTs) must undergo an independent bias audit within one year of use, with audit information made public and required notices provided. That audit is only one compliance step: employers still have to meet other antidiscrimination and disability-access obligations wherever they operate.
What does it mean to audit an AI hiring tool?
An AI hiring audit is a structured review of how an automated system affects candidates and whether its inputs, performance, and outcomes create unjustified disadvantages for protected or otherwise relevant groups. The review should cover the complete decision path—not just the model’s code or a vendor’s marketing claims.
As an Amazon Associate I earn from qualifying purchases.
Free tools Windows power users keep installed
One-click scans. No signup required.
Start by identifying what the product actually does. Common examples include resume screening, candidate ranking, skills assessments, interview recommendations, and video, speech, or voice analysis. A product does not become legally covered—or escape coverage—simply because its supplier calls it “AI.” The important questions are which employment decision it influences, how much weight its recommendation receives, and which jurisdiction governs that use.
New York City’s Department of Consumer and Worker Protection (DCWP) describes an AEDT as a system using machine learning, statistical modeling, data analytics, or artificial intelligence that substantially assists or replaces discretionary decision-making. Its AEDT FAQ is the starting point for determining whether a particular tool and use are covered by Local Law 144.
Popular on Amazon
- 【90 Degree Full Function USB C Adapter】Supports the latest USB4 standard for up to 40Gbps of bandwidth, 8K video support, and up to 240W Fast charging; Support USB PD 3.1 Extended Power Range (EPR) including 140W, 180W and 240W. Your productivity is easily doubled. *Warm Tip: For the best fit, please make sure the USB-C port is fully exposed; thick or bulky cases may affect compatibility.
- 【Reliable Quality】Made from high-quality, scratch-resistant aluminum alloy, delicate metal frame. get excellent wear resistance and heat dissipation. Non-Slip Design for easier insertion and removal.
- 【Small and portable】Ultra thin size, ultra compact design, a petite physique that matches the minimalist design of your modern laptops. Coming with a slim figure that won't block the port next to the one it's plugged into.
- 【Compatibility】Widely work with most usb c, thunderbolt 4/3 devices. Compatibilit for ROG Ally, Steam Deck, Switch, MacBook Pro, Quest 3 VR headset, Surface Book 2, iPhone 18/17/16/15, Samsung Galaxy S23/S22/S21, Note 20/10, Google Pixel, ChromeBook, etc.
- 【Excellence Customer Support】If you experience any issues or cable failure, simply reach out to our dedicated customer service team. We provide a prompt response within 18 hours, providing you with the necessary technical support to resolve any concerns.
How do I know whether our hiring software is covered?
Coverage depends on the tool’s function, the decision it affects, and where candidates work or are considered for work. A human remaining in the process does not automatically take a tool outside the rules if the software substantially shapes who advances.
- List every automated system. Include applicant-tracking filters, ranking models, assessments, interview-scoring tools, and systems that recommend rejection, progression, or hiring.
- Describe the decision. Record whether the output screens applicants, assigns a score, orders a shortlist, recommends an interview, or informs a final decision.
- Map geography. Note the job location, recruiting location, and any jurisdiction in which the decision has legal effect. Local Law 144 applies to covered employment decisions in New York City; other jurisdictions may impose different duties.
- Measure influence. Document whether recruiters routinely follow the recommendation, whether overrides are possible, and whether the output is a required gate.
- Check the current rule. Read the applicable regulator’s current definition, rules, FAQ, and enforcement material before relying on an internal classification.
For NYC use, the DCWP overview says an employer or employment agency may not use a covered AEDT unless a bias audit has been completed within the required period and required notices have been provided. This is a jurisdiction-specific summary, not a complete survey of federal, state, and local requirements.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallCrashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWhat must a New York City AEDT audit include?
For a covered NYC tool, an employer or employment agency must ensure an impartial evaluation by an independent auditor within one year of use, make the audit information publicly available, and provide the notices required by the law. The DCWP FAQ defines a bias audit as “an impartial evaluation by an independent auditor.”
At minimum, the audit calculates selection or scoring rates and impact ratios for:
- sex categories;
- race and ethnicity categories; and
- intersectional categories combining sex with race or ethnicity.
A selection rate is the share of people in a group who are selected or advance at the measured stage. A scoring rate applies the same idea to a defined score or score band. An impact ratio compares one group’s rate with a comparison group’s rate. The auditor must state how each category, outcome, comparison group, and population was defined; the figures are not meaningful without those choices.
Also popular on Amazon
- 5-in-1 Connectivity: Equipped with a 4K HDMI port, a 5 Gbps USB-C data port, two 5 Gbps USB-A ports, and a USB C 100W PD-IN port. Note: The USB C 100W PD-IN port supports only charging and does not support data transfer devices such as headphones or speakers.
- Powerful Pass-Through Charging: Supports up to 85W pass-through charging so you can power up your laptop while you use the hub. Note: Pass-through charging requires a charger (not included). Note: To achieve full power for iPad, we recommend using a 45W wall charger.
- Transfer Files in Seconds: Move files to and from your laptop at speeds of up to 5 Gbps via the USB-C and USB-A data ports. Note: The USB C 5Gbps Data port does not support video output.
- HD Display: Connect to the HDMI port to stream or mirror content to an external monitor in resolutions of up to 4K@30Hz. Note: The USB-C ports do not support video output.
- What You Get: Anker 332 USB-C Hub (5-in-1), welcome guide, our worry-free 18-month warranty, and friendly customer service.
Passing an audit is not a universal certification that a system is fair, accurate, accessible, or lawful in every job. DCWP does not prescribe one specific action an employer must take after an audit result. Employers remain responsible for complying with all applicable antidiscrimination laws, even when an audit has been completed.
How should an employer map the hiring decision pipeline?
Build a written inventory before examining results. The goal is to show where a candidate can gain or lose an opportunity and whether the software or a person made the operative decision.
| Record | What to capture | Why it matters |
|---|---|---|
| Tool and version | Vendor, product name, model or ruleset version, configuration, and change dates | Results can change after a model or setting changes. |
| Job and location | Role, requisition, work location, recruiting market, and dates covered | Coverage and the relevant candidate population are jurisdiction- and role-dependent. |
| Decision stage | Application screen, assessment, ranking, interview invitation, or final recommendation | Disparity can arise at one gate and disappear or worsen later. |
| Inputs and output | Data fields, modalities, score or rank, threshold, and downstream action | Connects a model signal to the employment opportunity it affects. |
| Human involvement | Reviewer, override rules, override frequency, and whether staff see the automated score | Separates a recommendation from the actual decision and reveals rubber-stamping. |
Keep separate records for the tool’s recommendation and the final human outcome. If recruiters can override a score but rarely do, the practical effect may still be substantial assistance. If the tool is used for several jobs or populations, treat each materially different use as a separate audit question unless the auditor can justify combining them.
Which audit questions should be set before looking at results?
Define the question, outcome, population, and comparison method in advance. Otherwise, an organization can unintentionally select the most favorable slice of data after seeing the results.
- Outcome: Are you measuring selection, a score band, ranking position, progression to interview, or another concrete opportunity?
- Stage: At which gate can the tool change a candidate’s chance of continuing?
- Population: Which applicants were eligible, complete, and actually processed during the period?
- Groups: Which sex, race, ethnicity, intersectional, disability, language, or other groups are relevant to the job and modality, subject to lawful data handling?
- Comparison: What reference group or baseline will be used, and why is it appropriate?
- Time period: Which model version, requisitions, and dates are included?
- Missing data: How are unknown, undisclosed, or small groups handled without hiding a potentially important result?
There is no single metric that answers every fairness question. Testimony recorded in an EEOC hearing transcript noted a lack of consensus about which metrics and data auditors should use and how assumptions should be disclosed. That uncertainty is a reason to publish the method and its limitations, not to present one ratio as proof of compliance.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsAlso popular on Amazon
- Sleek 7-in-1 USB-C Hub: Features an HDMI port, two USB-A 3.0 ports, and a USB-C data port, each providing 5Gbps transfer speeds. It also includes a USB-C PD input port for charging up to 100W and dual SD and TF card slots, all in a compact design.
- Flawless 4K@60Hz Video with HDMI: Delivers exceptional clarity and smoothness with its 4K@60Hz HDMI port, making it ideal for high-definition presentations and entertainment. (Note: Only the HDMI port supports video projection; the USB-C port is for data transfer only.)
- Double Up on Efficiency: The two USB-A 3.0 ports and a USB-C port support a fast 5Gbps data rate, significantly boosting your transfer speeds and improving productivity.
- Fast and Reliable 85W Charging: Offers high-capacity, speedy charging for laptops up to 85W, so you spend less time tethered to an outlet and more time being productive.
- What You Get: Anker USB-C Hub (7-in-1), welcome guide, 18-month warranty, and our friendly customer service.
How should outcomes be checked across hiring stages?
Calculate the required NYC rates and impact ratios, then extend the review to every consequential stage. A single final-hire comparison can miss a disparity created during an application screen or assessment.
- Initial screen: Compare who is accepted for further review, not just who is eventually hired.
- Assessment: Examine score distributions, completion rates, and progression by group.
- Interview selection: Check whether recommendations or thresholds change the chance of an interview.
- Ranking and shortlist: Review both the composition of the shortlist and where groups appear in the ranking.
- Final decision: Separate the automated contribution from later human decisions and record overrides.
Interpret differences in context. A disparity can originate in an input field, a job-related requirement, a threshold, missing data, a technical failure, or a human practice around the output. Do not turn a rule-of-thumb threshold into proof that a tool is lawful, and do not treat a small or incomplete population as conclusive without explaining the limitation.
How do accuracy and accessibility fit into a bias audit?
Fair group rates are not enough if the system misunderstands applicants or prevents some people from participating. Review whether each input is job-related, whether the tool functions consistently for the people expected to use it, and whether an applicant can obtain a reasonable adjustment.
The UK government’s Responsible AI in Recruitment guidance highlights potential group-accuracy problems in face, speech, and voice analysis. For these modalities, ask:
Recommended Free Tools
- Does the system recognize differences in speech, accent, voice, facial movement, lighting, or communication style that are unrelated to job performance?
- Can an applicant complete the process without a camera, microphone, timed response, or other modality that creates a barrier?
- Is there a clearly communicated route to request a reasonable adjustment or an alternative assessment?
- Are accommodation requests recorded separately from performance outcomes and handled confidentially?
- Were the validation population and test conditions similar to the actual applicant population and use?
Accessibility duties vary by jurisdiction, and the UK guidance is not a substitute for local legal advice. Its practical warning is broadly useful: an audit should examine the candidate experience and error patterns, not only aggregate selection ratios.
Also popular on Amazon
- Portable and powerful USB-C HUB: BENFEI USB Type-C HUB, with super-soft and knot-free silicone woven design cable, meets most mobile office needs. Compact, lightweight, stylish, and powerful portable USB C Hub equipped with 1 x HDMI port, 1 x 100W charging, and 3 x USB ports. 18-month warranty, 24-hour response, to ensure you feel at ease when using our product.
- Design centered on comfort and reliability: Thanks to BENFEI's end-to-end in-house cable production capability, in-house PCBA and assembly capability, using the industry's most advanced silicone woven design and process, 20cm cable in length, no knots, super-soft, the HUB is easy to use in all scenarios: laptop, tablet, stand etc. Super-soft, 25000+ life cycles, to meet your daily carrying and office needs.
- 100W Charging: Support up to 90W USB C pass-through charging via Type-C port to keep your laptop powered. 10W is reserved for other interface operations. No data and video function on the Type-C port.
- 4K HDMI Display: The HDMI port supports media display at resolutions up to 4K 30Hz, keeping every incredible moment detailed and ultra vivid. Please note that the C port of the Host device needs to support video output.
- Transfer Files in Seconds: Transfer files and from your laptop at speeds up to 10 Gbps with USB A 3.2 port. Extra 2 USB A 2.0 ports are perfectly for your keyboards and mouse.
What should employers request from vendors?
Vendor documentation is useful evidence, but it does not transfer the employer’s legal responsibility. Request enough information to understand the system’s intended use and limitations without assuming that a supplier’s validation claim proves suitability for your jobs.
- Intended and prohibited uses, including whether the product is designed to rank, filter, score, or recommend.
- All input fields, derived features, thresholds, and configurable settings that your organization enabled.
- Validation population, sample composition, outcome definition, and known limitations.
- Performance or error information for the modalities used, including face, speech, or voice analysis where applicable.
- Model, ruleset, and configuration version history, with dates of material changes.
- Available group-level outcome data and the supplier’s definitions for categories and missing values.
- Instructions for accessibility, accommodations, human review, and escalation.
- Retention, deletion, security, and access rules for candidate data.
Put requests and responses in the audit record. If a vendor will not disclose a material input, limitation, or version change, record that as a governance risk and consider whether the tool can be responsibly used for the decision.
What records should an employer keep?
Preserve a reproducible account of what was used, what was measured, and what was done. At a minimum, retain:
- the tool, model, ruleset, and configuration versions;
- jobs, locations, dates, candidate population, and data definitions;
- the audit question, metrics, comparison groups, assumptions, and missing-data treatment;
- selection or scoring rates and impact ratios, including intersectional results where required;
- accuracy, accessibility, accommodation, and error findings;
- vendor materials and communications;
- public notices and candidate notices;
- reviewer decisions, overrides, and escalation records; and
- remediation decisions, owners, dates, and follow-up checks.
The sources support documenting methods and assumptions but do not establish one universal documentation template. Make the record understandable to an independent reviewer: define every denominator, explain excluded cases, and identify which findings are statistically or operationally limited.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What should an employer do when an audit flags a disparity?
Do not dismiss the result because the vendor calls the system neutral, and do not assume that an audit finding automatically identifies the cause. First verify that the population, metric, and comparison are appropriate; then investigate the mechanism and the employment requirement involved.
Also popular on Amazon
- 【Read before You Buy】 Note: This adapter is USB 2.0 version and cannot transmit HDM, VGA, or other video signals. Nevertheless, they are fully capable of charging devices,duplicating files, and can be seamlessly integrated with USB-C earphones, wireless network adapters, and SD card adapters.
- 【Easy to Use and Safe】Thanks to modern manufacturing techniques, we've made these adapters sufficiently compact so that they can remain inserted in your device without noticeable protrusion or inconvenience. This means you don't have to constantly plug and unplug them, reducing the likelihood of wear and misalignment, and making them more convenient to carry around.
- 【Reliable Quality】Engineered for enhanced durability, durable aluminum alloy shell material,get excellent wear resistance and heat dissipation,the male USB C and female USB connectors & the male USB and female USB C connectors are designed to be plugged and unplugged up to 20000 times. Provides data transfer, device charging, memory space expansion, and additional computer peripherals.
- 【Think more for you】Do you need to connect a USB-C-enabled device to an odd device that only has USB-A ports? Then these USB-A male to USB-C female adapters are made for you.Simply insert the adapter into the USB-A port on the device, and you can instantly connect your USB-C device to the device that originally didn't have the correct connection port.
- 【Trusted xiwxi Quality and Service】As a responsible brand seller, you will receive xiwxi's 12-month Quality assurance and lifetime friendly customer service, you're not just buying a product,you're investing in peace of mind with all-round quality and exceptional customer service.
- Validate the finding. Check data quality, coding, missing categories, model version, and whether the result is confined to one job, stage, or time period.
- Trace the cause. Examine inputs, thresholds, training or validation limitations, workflow rules, and human reliance on the output.
- Assess job relevance. Ask whether the signal measures a genuine requirement or a proxy unrelated to successful performance.
- Protect candidates while reviewing. Add qualified human review, pause a gate, offer an alternative assessment, or stop using the affected feature when appropriate.
- Choose a remedy. Adjust the configuration, redesign the process, supplement the tool, replace it, or discontinue the use.
- Document and re-evaluate. Record the decision, responsible owner, effective date, candidate impact, and a plan to check the revised process.
NYC DCWP says Local Law 144 does not dictate a specific action based on audit results. That flexibility is not permission to ignore a problem: the New York State Comptroller’s 2026 report describes federal, state, and local discrimination protections as continuing to apply.
How can an audit avoid common failure modes?
- Auditing the product instead of the use: Scope the actual job, location, configuration, and decision stage.
- Relying on a vendor certificate: Obtain underlying definitions, populations, limitations, and version information.
- Looking only at final hires: Analyze every gate where candidates can be screened out.
- Using one aggregate number: Include required intersectional categories and investigate meaningful stage-specific differences.
- Ignoring accessibility: Test the modality and provide an adjustment path before candidates are excluded.
- Hiding assumptions: Publish or retain clear explanations of denominators, comparison groups, missing data, and exclusions.
- Treating completion as clearance: An audit is an evaluation, not a blanket finding that the system is lawful or fit for every role.
- Failing to monitor changes: Revisit the analysis after a model, threshold, job requirement, workflow, or candidate population changes.
What is a defensible audit program?
A defensible program combines legal scoping, independent review where required, outcome analysis, accessibility checks, vendor due diligence, and documented remediation. Begin with an inventory, define the question before reviewing results, measure the required NYC categories when applicable, and add stage-level and modality-specific checks that reflect the real candidate experience.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →For NYC employers, schedule the independent audit within one year of covered use, make the required information public, and deliver notices on the DCWP timetable. For employers elsewhere, identify the rules that apply to the work location and decision, then use the same evidence discipline: know what the tool does, measure who is affected, preserve assumptions, and act when the process creates an unjustified barrier.
Frequently Asked Questions
Does keeping a human reviewer make an AI hiring tool exempt from bias rules?
Not necessarily. Coverage can depend on whether the system substantially assists or replaces discretionary decision-making, even when a person remains in the workflow. Document how often reviewers follow or override the output and check the current rule for the jurisdiction and use.
Can an employer use an audit performed for another job or model version?
Only if the auditor can justify that the decision, configuration, population, time period, and model version are materially comparable. A changed threshold, job, workflow, or model can create a different audit question.
What if applicants do not disclose demographic information?
Do not silently treat unknown information as a neutral group. Define how undisclosed or missing data will be handled, report the limitation, and assess whether the remaining population is sufficient for a meaningful comparison while following privacy and data-protection requirements.
Are impact ratios the same as proof of discrimination?
No. They are indicators calculated from defined selection or scoring rates. Interpretation requires context about the population, job requirements, data quality, accuracy, accessibility, and other applicable legal standards.

