One Month until California’s ADS Regulations Come into Effect: Are You Ready?

On 1 October 2025, modifications to California’s employment regulations to address automated-decision systems (ADS) come into effect. These modifications are the result of a several-year process and multiple comment periods, with the final modifications approved in June.

The regulations prohibit unlawful employment discrimination in California. The purpose of the modifications is to explicitly specify that these non-discrimination requirements also apply to ADSs, driven by concerns that the use of ADSs may exacerbate existing biases and result in discriminatory outcomes.

With only a month to go until the modified regulations come into effect, in this blog post, we outline what you need to know and how to prepare.

How do California’s employment regulations define an ADS?

An automated-decision system (ADS) is defined as a computational process that is used to make a decision or facilitate human decision-making about an employment benefit. Benefits include hiring, employment, promotions, selection for training programs, discharge from employment or training, compensation, provision of a discrimination-free workplace, and freedom from disbarment.

ADSs may be powered by artificial intelligence (AI), machine learning algorithms, statistics, or other data processing techniques. They can perform a variety of tasks, such as using assessments or tests to make predictions about an applicant or employee, measure their skills, dexterity, reaction time, or other abilities, or measure personality, aptitude, attitude, or culture fit.

ADSs can also be used to analyze facial expressions, word choice, or voice in interviews or to analyze data acquired from a third party. They may also screen or categorize applicants or direct job adverts to targeted groups.

The definition also explicitly excludes several technologies that do not make a decision about employee benefits. This includes word processing software, spreadsheet software, data storage, spam- and robocall-filtering, and databases.

How are the technologies underpinning ADSs defined?

An algorithm is defined as a set of rules or instructions that are used by a computer to carry out calculations or other problem-solving. This includes using patterns detected in datasets to automate decision-making.

Artificial intelligence is defined as a machine-based system that uses inferences from its inputs to generate outputs such as predictions, content, recommendations, or decisions.

Finally, machine learning refers to the ability of a computer to automatically apply its learnings from its own analysis of inputs or experience to future calculations or tasks.

How are California’s employment regulations updated?

The principles of employment discrimination under California’s regulations are updated to specifically state that it is unlawful for an employer or another covered entity (such as an employment agency or an agent of an employer) to use an ADS or selection criteria that discriminates against an applicant or class of applicants or employees on a protected attribute.

The regulations also note that ADSs that measure skill, dexterity, reaction time, or other characteristics may result in discrimination against individuals with disabilities and that a reasonable accommodation may need to be provided to avoid unlawful discrimination. The same is noted for ADSs that analyze tone of voice, facial expressions, or other physical characteristics or behavior.

Moreover, the modifications explicate that certain specific prohibitions apply to discrimination resulting, at least in part, from the use of an ADS. This includes discrimination based on:

However, the regulations do note that discrimination may not be considered unlawful if the characteristic is job-related.

How are the record-keeping requirements updated?

Previously, employers were required to keep any employment records made or kept about an applicant or employee’s benefits for two years. The modifications extended this for four years, where the records that must be kept include applications, personnel records, referral records, selection criteria, and data entered into or produced by an ADS (ADS data).

What is the role of bias audits?

The regulations state that evidence of anti-bias testing may contribute to the defense of a claim of employment discrimination, where the quality, efficacy, recency, and scope of the anti-bias testing may influence how strong the defense is.

In other words, a comprehensive yearly bias audit to detect and mitigate bias could provide a defense to claims of unlawful discrimination.

How to prepare for California’s ADS regulations

With only a month to go until California’s modified employment regulations go into effect, acting fast is vital. Although existing equal opportunity legislation always applies to employment decisions made using automated tools, California’s modifications highlight the importance of comprehensive bias audits of automated tools.

Although the regulations apply to employers, vendors can play a key role in helping their clients to stay compliant.

If you are an employer:

• Create an inventory of all of the HR tech tools you use across the company
• Determine whether any meet the definition of an ADS
• Request documentation on any testing already carried out by the vendor of any tools not developed in-house
• Strengthen anti-bias evidence by carrying out a bias audit using real-life data, where possible, and document the data audited and how the audit was carried out

If you are a vendor:

• Create an inventory of all of the ADS you provide
• Ensure you have clear documentation on how your ADSs were designed, developed, and validated
• Regularly test tools for bias using real-life data, where possible, and document the results

Bias audits with Holistic AI

Whether you are a vendor of HR tech tools or an employer leveraging them, regular bias audits can reduce legal, reputational, and financial risks. With other laws already requiring bias audits of automated employment decision tools, they are fast becoming the standard for demonstrating the legality and fairness of HR tech tools.

Holistic AI is the leading provider of bias audits of automated-decision systems with rich, multidisciplinary expertise in I-O psychology, computer science, policy, and ethics. Get in touch with our experts to find out more about how Holistic AI can help you prepare for California’s ADS regulations.