Artificial intelligence is changing how California employers recruit, screen, and evaluate workers. Companies now use software to review resumes, rank applicants, score assessments, and support hiring decisions.
These tools can save time. However, they can also create legal problems when their results disadvantage workers based on protected characteristics.
California has responded with new rules for automated decision systems in employment. The regulations took effect on October 1, 2025. As a result, employers using AI hiring tools in 2026 should review how those systems work and how they affect applicants.
The rules do not ban artificial intelligence in hiring. Instead, they clarify how existing California employment discrimination laws apply when employers use automated systems.
Job applicants should also understand these changes. A hiring decision does not become exempt from discrimination law simply because software helped make it.
This article explains the main issues in simple terms. It provides general information and does not replace legal advice about a specific employment dispute.
How California AI Hiring Rules Affect Employment Decisions
California’s regulations focus on automated decision systems used in employment. These systems may rely on algorithms, machine learning, artificial intelligence, or other automated methods.
An employer may use these tools during recruitment, screening, hiring, promotion, or other employment decisions. The technology can assist a person or make recommendations before someone makes the final decision.
California law still requires employers to follow workplace discrimination rules when they use these systems. Employers cannot avoid those duties by relying on software supplied by another company.
Automated Hiring Tools Can Create Discrimination Risks
The California Civil Rights Department explains that automated systems may violate state law when they harm applicants or employees because of a protected characteristic.
Protected characteristics can include race, sex, gender, disability, religion, national origin, age, and other categories covered by California law.
The risk may not always be obvious. A tool can appear neutral while producing different results for different groups.
For example, software might rank applicants based on patterns found in past hiring data. If that historical data reflects unequal practices, the system may repeat or strengthen those patterns.
Another tool might evaluate how someone speaks during a video interview. That process could create problems if it treats accents, disabilities, speech patterns, or other characteristics differently.
Employers should therefore review both the purpose and the results of an automated system. They should not rely only on a vendor’s statement that a tool is fair.
What Counts as an Automated Decision System?
California’s regulations define terms related to automated decision systems, artificial intelligence, algorithms, machine learning, and automated-decision data.
The rules can reach more than software that makes a final hiring decision. A system may still matter when it helps screen, rank, recommend, or evaluate applicants.
Common examples may include resume screening software, applicant ranking systems, online assessments, and automated interview tools.
Some systems analyze test results or game-based assessments. Others may review voice, facial expressions, behavior, or other information.
The legal question depends on how the employer uses the system. Employers should document what each tool does and where it affects the employment process.
A clear inventory can help. List every automated system used during recruiting, hiring, promotion, performance review, and other employment decisions.
Recordkeeping and Vendor Oversight Matter
California’s rules also expand the importance of employment records. Covered employers and other entities must maintain certain employment records for at least four years.
That requirement includes automated-decision system data. Employers should review how they preserve information created during automated screening and selection.
Vendor relationships deserve attention as well. Many companies do not build their own hiring systems. Instead, they buy or license tools from third-party providers.
Using an outside provider does not remove the employer’s need to understand the tool. Employers should ask vendors how the system evaluates applicants and what information it uses.
Companies should also ask whether the vendor has tested the tool for unfair results. Contract terms should address access to records, system changes, security, and compliance support.
Employers can review additional compliance information through the site’s Legal Resources. The site’s guide to California privacy rules in 2026 also explains related requirements for automated decisionmaking technology.
AI Screening Can Raise Disability and Medical Inquiry Issues
Automated hiring tools can create special concerns for applicants with disabilities. A screening method may measure characteristics that do not accurately reflect a person’s ability to perform the job.
California’s regulations address this risk directly. For example, automated assessments may measure reaction time, dexterity, or other abilities.
Those tests can disadvantage some applicants with disabilities. An employer may need to provide reasonable accommodation so the applicant has an equal opportunity.
Video interview systems can create similar concerns. A tool that evaluates facial expressions, voice, or physical behavior may work differently for people with certain disabilities.
Employers should not assume that an automated score tells the full story. They should consider whether the tool creates barriers that relate to disability rather than actual job ability.
Applicants May Need Reasonable Accommodation
Employers should have a clear process for accommodation requests during automated screening. Applicants should also know how to request help when a tool creates an accessibility problem.
For example, an applicant may have difficulty using a timed assessment because of a disability. Another applicant may need an alternative to a video-based evaluation.
California regulations state that employers may need reasonable accommodations when automated systems affect applicants with disabilities or certain religious needs.
Automated assessments can also create medical inquiry problems. A test, question, or puzzle that reveals disability information may cross legal boundaries before a job offer.
Employers should review questions and assessments before using them. The fact that software asks the question does not make the inquiry lawful.
Job applicants should save emails, assessment instructions, accommodation requests, and responses. Those records may help explain what happened if a dispute develops.
What Employers and Job Applicants Should Do in 2026
California employers should treat 2026 as an important compliance year. The employment discrimination regulations already apply, while separate privacy duties for certain automated decisions arrive in 2027.
The California Privacy Protection Agency has adopted rules for automated decisionmaking technology under the CCPA. Covered businesses using ADMT for significant decisions must comply with those requirements beginning January 1, 2027.
Employment is one area covered by the significant-decision rules. Depending on the business and the technology, consumers may receive notice and certain rights related to ADMT.
Employers should therefore coordinate employment, privacy, human resources, and technology reviews. A tool can raise more than one legal issue at the same time.
Build a Practical AI Hiring Compliance Process
Start by identifying every automated tool used in the employment process. Do not limit the review to systems labeled as artificial intelligence.
Next, document what each tool does. Record the information it receives, the result it produces, and the decision it helps support.
Employers should then review whether the system creates different outcomes for protected groups. Testing may help identify problems before they affect more applicants.
Human review is also important. Managers should understand that a software recommendation should not automatically replace careful decision-making.
Training can reduce mistakes. Recruiters and hiring managers should know how automated tools work and when an applicant may need an accommodation.
2026 Checklist for Employers and Applicants
Employers can begin with several practical steps:
- List all automated hiring and employment tools.
- Identify what data each system collects.
- Review how each tool affects hiring decisions.
- Check for possible discrimination or adverse impact.
- Create a process for reasonable accommodation requests.
- Review vendor testing and contract terms.
- Preserve required automated-decision records.
- Train recruiters and hiring managers.
- Review privacy requirements that begin in 2027.
- Reassess tools when vendors change their systems.
Job applicants can take practical steps too. Read assessment instructions carefully and save important communications.
If a disability affects your ability to use a hiring tool, consider asking about a reasonable accommodation. Keep a copy of the request and the employer’s response.
If you believe an automated system affected you because of a protected characteristic, document the facts. Save job postings, emails, test instructions, rejection notices, and other relevant records.
The California Civil Rights Department provides official information about the state’s regulations and employment discrimination protections.
Read the California Civil Rights Department’s guidance on automated employment decision systems.
Visitors can also review the site’s Practice Areas for additional information about available legal services. The Legal Resources section offers more general California legal guidance.
Artificial intelligence will likely remain part of modern recruiting. The legal focus is therefore shifting toward how companies design, select, test, and use these systems.
California employers should not wait for a complaint before reviewing their tools. Early testing, clear documentation, and effective training can help identify risks sooner.
Applicants should remember that automated hiring does not remove workplace protections. Existing discrimination rules still apply when technology plays a role in an employment decision.
As California continues to develop rules for automated decisions, employers should monitor guidance from the Civil Rights Department and the California Privacy Protection Agency. A regular review can help keep hiring practices aligned with changing requirements.
