EEOC Issues Guidance on Employer Use of AI Decision-Making Tools

On May 12, 2022, the Equal Employment Opportunity Commission (EEOC) issued pointers cautioning employers on the use of digital software program, algorithms, and synthetic intelligence (AI) to make employment-related choices. The pointers highlighted areas that will increase potential Americans with Disabilities Act (ADA) violations. The EEOC suggestions are the primary substantive output of the EEOC’s Artificial Intelligence and Algorithmic Fairness Initiative, launched in October 2021, which focuses on addressing the rising use by employers of rising AI applied sciences.
Background
Employers now have entry to all kinds of computer-based instruments to help in a variety of actions, from hiring workers to monitoring work efficiency and figuring out pay and promotions. Common AI instruments embrace:

Algorithm-based resume scanners and software program that prioritize purposes utilizing sure key phrases. These instruments can charge job candidates on the idea of key elements.
Virtual assistants or “chatbots” that may ask job candidates about their {qualifications}.
Video interviewing software program that may consider candidates based mostly on their facial expressions and speech patterns.
Testing software program that may present “job match” scores.

Although many employers have carried out these instruments in an effort to save lots of money and time, the EEOC cautions that the instruments might drawback job candidates and workers with disabilities.
Potential ADA Violations
The EEOC’s steerage describes three attainable methods by which an employer’s AI use might violate the ADA:

An employer’s failure to adjust to the responsibility to offer affordable lodging (corresponding to different testing codecs and prolonged time) as a result of algorithmic evaluations of candidates and workers;
An employer’s use of algorithmic decision-making instruments that deliberately or unintentionally display screen out certified people and discriminate in opposition to these with disabilities who’re in a position to do the job with affordable lodging; and
An employer’s use of instruments that fail to adjust to the ADA’s restrictions on disability-related inquiries and medical examinations.

The steerage clearly states that employers might not rely on a third-party vendor’s illustration that an AI software is “bias free” for validation functions.
Recommendations for Employers
The EEOC’s steerage recommends employers implement the next practices when utilizing software program, algorithms, and AI instruments:

1) Train workers to acknowledge and promptly course of requests for affordable lodging.
2) Develop alternate means of ranking job candidates.
3) If an algorithmic decision-making software is run by a 3rd celebration, instruct the entity to promptly ahead all requests for lodging or enter into an settlement for the third celebration to offer affordable lodging.
4) Use algorithmic decision-making instruments which were designed to be accessible to people with disabilities.
5) Inform all job candidates and workers that affordable lodging can be found and supply clear and accessible directions for requesting such lodging.
6) Clearly describe the traits the algorithm is designed to entry, the strategy by which these traits are assessed, and the variables and elements that will have an effect on the ranking.
7) Ensure that the algorithmic decision-making instruments solely measure skills or {qualifications} which are actually crucial for the job.
8) Ensure that such skills and {qualifications} are measured immediately, somewhat than by means of traits or scores which are correlated with these skills or {qualifications}.
9) Confirm that the software doesn’t ask questions more likely to elicit details about a incapacity or search details about a person’s bodily or psychological impairment or well being, until such inquiries are associated to a request for affordable lodging.

In addition to the EEOC’s steerage, comparable laws concerning AI decision-making instruments are being proposed by different regulatory our bodies, together with California’s Fair Employment and Housing Council. DWT’s latest advisory on California’s draft laws may be discovered right here.
The EEOC will increase its steerage to deal with additional developments within the use of AI. DWT will proceed to offer updates on this rising intersection of AI expertise and employment legislation. In the meantime, in case you have any questions concerning the implementation of these newly advisable practices, please be happy to contact DWT’s Employment Services Group.

https://www.lexology.com/library/detail.aspx?g=e26e3c65-b766-4a17-8905-541aa6369040

Recommended For You