California Moves to Ban AI Employment Discrimination

California Moves to Ban AI Employment Discrimination

Sacramento, CAOn June 2, California’s Automated Decisions Safety Act (AB 1018), passed the Assembly and moved on to Senate committee consideration. The bill is designed to protect employees, applicants for employment and consumers from AI-driven discrimination banned by federal law and the California Labor Code. The problem is also at the heart of Mobley v. Workday, Inc., a class action lawsuit in the Northern District of California.

AB 1018 is now part of a multi-front effort to address the invisible bias that can be buried in an algorithm. This issue has been raised and shot down a few times. Maybe this time now?

What’s the problem with automated decision-making systems?

The use of automated decision systems (ADS) has become part of Californian’s daily lives – in areas such as housing, employment, and even in criminal justice sentencing and probation decisions. The good news is that ADS can make decision making faster and may produce more consistent results.

However, according to the bill’s author, the algorithms that power these systems are often vulnerable to issues such as unrepresentative data, faulty classifications and flawed design. These shortcomings can result in biased, discriminatory, or unfair outcomes. Rather than solving systemic problems, poorly designed ADS can hurt the people they are meant to help.

Real people

Derek Mobley puts a human face on the problem.

Mobley is an African American man, over the age of forty with a degree from Morehouse College, an HBCU. He has a medical history of depression.

Beginning in 2017, he applied for more than one hundred positions with employers that exclusively use Workday, Inc. as a screening platform for hiring. Every application was rejected.

When Mobley applied to an online job ad on a third-party website such as LinkedIn, Indeed, Monster, or CareerBuilder, the site would direct him to the Workday platform. He would then upload a resume or enter his information manually and take a personality test.

Inevitably, he would get a rejection email. Sometimes the rejection took less than an hour or arrived in the middle of the night. Something seems wrong here.

Finally, he filed a federal employment discrimination lawsuit on behalf of himself and other similarly situated workers.

More common than many think

A 2023 EEOC technical advice clarified that employers may be held liable under Title VII of the Civil Rights Act for employment discrimination that results from the use of algorithmic decision-making tools. In the same year, the EEOC settled the first AI-driven discriminatory hiring lawsuit against iTutorGroup. The EEOC determined that the employer’s hiring software was programmed to reject female applicants age 55 and older and male applicants age 60 and older in violation of the Age Discrimination in Employment Act (ADEA). iTutorGroup was ultimately ordered to pay $365,000 to more than 200 potential tutors automatically rejected because of age.

Automated Decisions Safety Act – two steps and threefold impact

AB 1018 takes a two-step approach. It would impose immediate requirements on the developers of automated decision-making systems (like Workday). Beginning on January 1, 2027, the law would impose additional requirements on organizations using ADS (like employers). Together, these requirements will offer significant protections to workers and job applicants (like Derek Mobley).

Step one – developer mandates

When enacted, the law would require ADS developers to conduct and document an initial performance evaluation of the system’s compliance with anti-discrimination laws. The evaluation would have to be repeated annually in consultation with an independent third-party auditor. The ADS developer would also be required to disclose the results of these evaluations to those using the ADS to make important decisions about hiring, firing and promotions, among other issues.

Step two – employer mandates

Beginning on January 1, 2027, employers must notify any individual affected by an ADS-assisted decision to:

  • notify the individual about the use of the system;
  • give him or her the opportunity to opt out of the use of the ADS;
  • correct the individual’s incorrect personal information; and
  • appeal the outcome.

In addition, any employer using ADS must:

  • minimize collection, use, retention, and sharing of a subject’s personal information; and
  • if the decision-making system affects more than 5,999 people over a three-year period, the organization must contract with a third-party independent auditor to assess the impact of the system.

When fully implemented, the law would provide a “check-and-correct procedure” at two different and independent stages.

AB 1018 has some teeth

The bill would also authorize public prosecutors, the Civil Rights Department, and the California Labor Commissioner to bring civil actions to enforce violations of the bill’s requirements.

What should California workers and job applicants do?

At its most elemental level, the answer may be to act like Derek Mobley.

  • Wake up – Yes, automated employment application discrimination happens;
  • Speak up – There’s no reason to be abashed because of age, gender, race or health issues. It’s not just you. Others are experiencing the same problems; and
  • Stand up – Get some help from an experienced employment lawyer.

AB 1018, if enacted, would take a significant legislative step toward fixing a problem that courts have already recognized.

Google News Website Posting For Attorneys
Source link

Recommended For You

Leave a Reply

Your email address will not be published. Required fields are marked *

Home Privacy Policy Terms Of Use Anti Spam Policy Contact Us Affiliate Disclosure DMCA Earnings Disclaimer