• NYC Local Law July 5th, 2023, and aims to detect bias in hiring and employment decisions from automated employment decision tools (AEDTs)

    • Companies who use AEDTs covered under the law will need to perform independent audits on those tools and post the results to a public website

    • Proceptual can help make the process easy, and we have a structured methodology to help you comply quickly

NYC Local Law 144 (NYC LL 144) – Overview

NYC Local Law 144 (NYC LL 144) is one of the first laws regulating automated employment decision tools (AEDT).  The goal is to prevent bias in the hiring process.

NYC LL 144 specifies that it is regulating “Automated Employment Decision Tools”, and that employers using these tools to hire NYC residents will need to perform a third-party, independent audit bias for each of them.

AEDTs refer to the following:

Any computational process, derived from machine learning, statistical modeling, data analytics, or artificial intelligence, that issues simplified output, including a score, classification, or recommendation, that is used to substantially assist or replace discretionary decision making for making employment decisions that impact natural persons. The term “automated employment decision tool” does not include a tool that does not automate, support, substantially assist or replace discretionary decision-making processes and that does not materially impact natural persons, including, but not limited to, a junk email filter, firewall, antivirus software, calculator, spreadsheet, database, data set, or other compilation of data.

You will notice that neither artificial intelligence nor machine learning are specifically mentioned in this description.  While an AEDT very well could leverage AI or machine learning, it is not required to be considered one.

According to the text, a bias audit is an:

impartial evaluation by an independent auditor. Such bias audit shall include but not be limited to the testing of an automated employment decision tool to assess the tool’s disparate impact on…. [covered groups]

Employers must also notify applicants that these tools will be used beforehand, and once a bias audit is complete, the results must be posted publicly on the web.

Failure to comply could result in fines of up to $1,500 per day, per occurrence.

It is also important to note that the law is officially in effect as of January 1st, 2023, but enforcement has been postponed until July 5th, 2023.  Companies who are required to have bias audits but have yet to perform them are technically not in compliance, although the enforcement itself has been postponed.

How to become compliant

Proceptual makes compliance easy, breaking it down to several steps:


    • Is an audit necessary?

    • Which tools must be audited?


    • Collect data and clean

    • Comply with internal and external prviacy requirements


    • Produce Selection rate and Impact Ratio

    • Review data internally for accuracy


    • Publish audit results on custom Proceptual landing page and link back


    • Recommend notice requirements and certify their placement


How are these fines calculated?

There are two pieces of the fine that employers could be fined for:

    1. Using a non-audited AEDT

    1. Not providing notice to applicants or employees of the use of an AEDT

And each of these violations could be fined up to $1500 per day.

This means that an employer that is using two non-audited AEDTs and not providing notice could face fines of up to $4500 per day – that’s over $30,000 a week.

How do you see if a company is in compliance?

The posting of the audit results and the notices to candidates may both be public-facing, meaning it is relatively easy to look at an employer’s job board and see if they are in compliance.

For the audit results, they must be shared publicly – but not necessarily on the company’s site.  Part of Proceptual’s service, for example, is to post the results on a microsite that an employer can link back to.  This ensures the results are always posted and meeting regulatory requirements, but also that they can be quickly updated as the bias reports are updated every year.

For the notices to candidates, the following guidance was provided:

(1) Including notice on the careers or jobs section of its website in a clear and conspicuous the manner at least 10 business days prior to the use of an AEDT;

(2) Including notice in a job posting at least 10 business days prior to use of an AEDT; or,

(3) Providing notice to candidates for employment via U.S. mail or e-mail at least 10 business days prior to the use of an AEDT.

This means that there is a way to avoid public sharing of the notice if you choose to provide that notice to candidates via email or mail 10 days prior to the use of an AEDT – which could prolong the hiring process for employers who choose to go that route.

How long does compliance take?

It depends on a few factors:

    1. The method of performing the audit and posting the results

    1. The capacity of your third-party vendor to perform such work

    1. The number of AEDTs an organization is using

    1. The state of the data, and how much standardization work needs to be done

First, as the bias audit is a statistical data analysis, there is a certain level of data science that will go into the effort.  That said, it doesn’t have to be a manual process.  Proceptual’s tech-driven approach automates and standardizes much of the process, but applies expertise to analyzing and making recommendations on the outputs.

Second, there could be tens of thousands of employers who will need to comply with this ai hiring law in a short amount of time. Depending on the third-party auditor you choose, you may be waiting in line.  Proceptual’s approach is fully scalable, so we do not anticipate there being a need for delay when working with us, but taking a fully manual approach could introduce delays.

Third, each AEDT needs to be audited separately. The more tools, the more audits. While Proceptual automates much of the process, there is a marginal time cost for each additional tool.
Last, the number of errors, any missing information, changes in format from one data set to the next, and other factors, could require additional work when preparing the data for analysis.
All in all, Proceptual anticipates a turnaround of fewer than 4 weeks for most of our customers – but we will provide a full timeline after understanding more about your needs.

Have further questions?  Reach out to Ken Hellberg and he will be happy to help:  ken@proceptual.com