by Arsen Kourinian, Ruth Zadikany, and Remy N. Merritt

Left to proper: Arsen Kourinian, Ruth Zadikany, and Remy N. Merritt (pictures courtesy of Mayer Brown)
The California Civil Rights Council (CRC) not too long ago introduced that it has finalized laws that make clear how California’s anti-discrimination legal guidelines apply to the usage of synthetic intelligence (Al) and automatic determination techniques (ADSs) in employment decision-making (the “Laws”). The Laws present that the usage of an ADS (together with Al) in making employment choices can violate California legislation if such instruments discriminate towards staff or candidates — both instantly or attributable to disparate affect — on the premise of protected traits (together with race, age, spiritual creed, nationwide origin, gender, and incapacity).
Efficient on October 1, 2025, the Laws amend the prevailing regulatory framework relevant to the California Truthful Employment and Housing Act (FEHA) and can apply to all employers in California that use “synthetic intelligence, machine-learning, algorithms, statistics, and/or different information processing” to facilitate human decision-making with respect to the recruitment, hiring, and promotion of job candidates or staff. In saying the issuance of the ultimate laws, the Civil Rights Council defined that, whereas “these instruments can deliver myriad advantages, they’ll additionally exacerbate present biases and contribute to discriminatory outcomes.” Consequently, the laws purpose to:
- Make it clear that the usage of an ADS might violate California legislation if it harms candidates or staff based mostly on protected traits, corresponding to gender, race, or incapacity.
- Guarantee employers and lined entities preserve employment information, together with automated determination information, for no less than 4 years.
- Affirm that ADS assessments, together with exams, questions, or puzzle video games that elicit details about a incapacity, might represent an illegal medical inquiry.
- Add definitions for key phrases used within the laws, corresponding to “automated-decision system” and “proxy.”
The laws broadly outline an ADS as any “computational course of that comes to a decision or facilitates human determination making relating to an employment profit” that “could also be derived from and/or use synthetic intelligence, machine-learning, algorithms, statistics, and/or different information processing methods.” 2 Cal. Code Regs. § 11008.1(a). The definition of ADS expressly consists of Al, which can be outlined broadly within the laws to incorporate “[a] machine-based system that infers, from the enter it receives, easy methods to generate outputs,” which might embrace “predications, content material, suggestions or choices.” § 11008.1(c).
The laws present illustrative examples of the varieties of duties ADS carry out, together with:
- Utilizing computer-based assessments or exams, corresponding to questions, puzzles, video games, or different challenges, to (i) make predicative assessments about an applicant or worker;(ii) measure an applicant’s or worker’s abilities, dexterity, response time, and/or different talents or traits; (iii) measure an applicant’s or worker’s character trait, aptitude, angle, and/or “cultural match;” and/or (iv) display, consider, categorize, and/or suggest candidates or staff;
- Directing job ads or different recruiting supplies to focused teams;
- Screening resumes for explicit phrases or patterns;
- Analyzing facial expressions, phrase alternative, and/or voice in on-line interviews; or
- Analyzing worker or applicate information acquired from third events.
The laws prohibit employers from utilizing an ADS or choice standards (together with a qualification commonplace, employment check, or proxy) that discriminates towards candidates or staff based mostly on protected classes outlined underneath the FEHA. § 11009(f). The time period “proxy” is newly outlined as “[a] attribute or class carefully correlated with” a protected class underneath FEHA. § 11008(1). The brand new laws additional clarify that the usage of facially impartial ADS choice instruments which have an “antagonistic affect” on candidates or staff based mostly on a protected attribute are impermissible underneath FEHA except the employer or lined entity can present that the selectin apply is “job-related and in step with enterprise necessity.” § 11017(e). Thus, as with different choice standards utilized in making employment choices, even when an employer doesn’t deliberately use an ADS (together with Al) to discriminate amongst candidates or staff, it may be chargeable for violating FEHA if the usage of the ADS creates disparate affect.
The laws warning that the usage of an ADS that, for instance, measures an applicant’s ability, dexterity, response time, and/or different talents or traits might discriminate towards people with sure incapacity or different protected traits. § 11016(c)(5). Equally, an ADS that, for instance, analyzes an applicant’s tone of voice, facial expressions, or different bodily traits or habits might discriminate towards staff or candidates based mostly on race, nationwide origin, gender, incapacity, or different protected traits. § 11016(d)(1). Accordingly, to keep away from illegal discrimination, employers might have to supply cheap lodging to an applicant or worker in step with the FEHA’s spiritual creed and incapacity protections. §§ 11016(c)(5) and (d)(1).
The Laws present that, to defend towards a discrimination declare based mostly on the usage of an ADS, employers can present that they carried out “anti-bias testing or related proactive efforts to keep away from illegal discrimination” previous to and after adopting an ADS. § 11009(f). The laws establish six related facets of such testing, together with the standard, efficacy, recency, and scope of such testing, in addition to the outcomes of the testing or different due diligence and the employer’s response to the outcomes (e.g., whether or not and the way the employer responded to the outcomes).
The laws have been additionally amended to require employers and lined entities to protect personnel and different employment information topic to the next necessities:
- Information have to be retained for a minimum of 4 years from the later of (a) the date the document was made, or (b) the date of the personnel motion — a rise from the earlier requirement to protect information for 2 years;
- Information topic to this requirement embrace choice standards, automated determination system information, functions, personnel information, membership information, employment referral information, and different information “created or acquired by the employer or different lined entity coping with any employment apply and affecting any employment advantage of any applicant or worker;” and
- “Automated-decision system information” consists of (a) any information utilized in or ensuing from the applying of an ADS, corresponding to information offered by or about particular person candidates or staff, or information reflecting employment determination or outcomes, and/or (b) any information used to develop or customise an ADS to be used by a selected employer or different lined entity.
The laws lengthen legal responsibility for ADS-driven discrimination to an employer’s “agent,” which is outlined as anybody “appearing on behalf of an employer, instantly or not directly, to train a operate historically exercised by the employer or another FEHA-regulated exercise,” corresponding to applicant recruitment, screening and hiring, promotion, or choices relating to pay, advantages, or go away, “together with when such actions and choices are performed in entire or partially via the usage of an automatic determination system.”
Employers that use ADS techniques in making employment choices and Al distributors whose merchandise are used within the employment area ought to take concrete steps to arrange for the implementation of the brand new Laws on October 1, together with:
- Figuring out the ADS techniques utilized in employment decision-making;
- Reviewing and amending document retention insurance policies to make sure information are retained for a minimum of 4 years;
- Performing anti-bias testing, establishing a plan outlining the frequency and nature of such testing, and documenting the testing course of/standards, outcomes, and steps taken to deal with outcomes, as applicable; and
- Updating anti-discrimination and cheap lodging insurance policies to deal with the usage of ADS techniques.
Arsen Kourinian and Ruth Zadikany are Companions and Remy N. Merritt is an Affiliate at Mayer Brown. This publish first appeared on the agency’s weblog.
The views, opinions and positions expressed inside all posts are these of the writer(s) alone and don’t signify these of the Program on Company Compliance and Enforcement (PCCE) or of the New York College College of Regulation. PCCE makes no representations as to the accuracy, completeness and validity or any statements made on this web site and won’t be liable any errors, omissions or representations. The copyright of this content material belongs to the writer(s) and any legal responsibility with reference to infringement of mental property rights stays with the writer(s).
by Arsen Kourinian, Ruth Zadikany, and Remy N. Merritt

Left to proper: Arsen Kourinian, Ruth Zadikany, and Remy N. Merritt (pictures courtesy of Mayer Brown)
The California Civil Rights Council (CRC) not too long ago introduced that it has finalized laws that make clear how California’s anti-discrimination legal guidelines apply to the usage of synthetic intelligence (Al) and automatic determination techniques (ADSs) in employment decision-making (the “Laws”). The Laws present that the usage of an ADS (together with Al) in making employment choices can violate California legislation if such instruments discriminate towards staff or candidates — both instantly or attributable to disparate affect — on the premise of protected traits (together with race, age, spiritual creed, nationwide origin, gender, and incapacity).
Efficient on October 1, 2025, the Laws amend the prevailing regulatory framework relevant to the California Truthful Employment and Housing Act (FEHA) and can apply to all employers in California that use “synthetic intelligence, machine-learning, algorithms, statistics, and/or different information processing” to facilitate human decision-making with respect to the recruitment, hiring, and promotion of job candidates or staff. In saying the issuance of the ultimate laws, the Civil Rights Council defined that, whereas “these instruments can deliver myriad advantages, they’ll additionally exacerbate present biases and contribute to discriminatory outcomes.” Consequently, the laws purpose to:
- Make it clear that the usage of an ADS might violate California legislation if it harms candidates or staff based mostly on protected traits, corresponding to gender, race, or incapacity.
- Guarantee employers and lined entities preserve employment information, together with automated determination information, for no less than 4 years.
- Affirm that ADS assessments, together with exams, questions, or puzzle video games that elicit details about a incapacity, might represent an illegal medical inquiry.
- Add definitions for key phrases used within the laws, corresponding to “automated-decision system” and “proxy.”
The laws broadly outline an ADS as any “computational course of that comes to a decision or facilitates human determination making relating to an employment profit” that “could also be derived from and/or use synthetic intelligence, machine-learning, algorithms, statistics, and/or different information processing methods.” 2 Cal. Code Regs. § 11008.1(a). The definition of ADS expressly consists of Al, which can be outlined broadly within the laws to incorporate “[a] machine-based system that infers, from the enter it receives, easy methods to generate outputs,” which might embrace “predications, content material, suggestions or choices.” § 11008.1(c).
The laws present illustrative examples of the varieties of duties ADS carry out, together with:
- Utilizing computer-based assessments or exams, corresponding to questions, puzzles, video games, or different challenges, to (i) make predicative assessments about an applicant or worker;(ii) measure an applicant’s or worker’s abilities, dexterity, response time, and/or different talents or traits; (iii) measure an applicant’s or worker’s character trait, aptitude, angle, and/or “cultural match;” and/or (iv) display, consider, categorize, and/or suggest candidates or staff;
- Directing job ads or different recruiting supplies to focused teams;
- Screening resumes for explicit phrases or patterns;
- Analyzing facial expressions, phrase alternative, and/or voice in on-line interviews; or
- Analyzing worker or applicate information acquired from third events.
The laws prohibit employers from utilizing an ADS or choice standards (together with a qualification commonplace, employment check, or proxy) that discriminates towards candidates or staff based mostly on protected classes outlined underneath the FEHA. § 11009(f). The time period “proxy” is newly outlined as “[a] attribute or class carefully correlated with” a protected class underneath FEHA. § 11008(1). The brand new laws additional clarify that the usage of facially impartial ADS choice instruments which have an “antagonistic affect” on candidates or staff based mostly on a protected attribute are impermissible underneath FEHA except the employer or lined entity can present that the selectin apply is “job-related and in step with enterprise necessity.” § 11017(e). Thus, as with different choice standards utilized in making employment choices, even when an employer doesn’t deliberately use an ADS (together with Al) to discriminate amongst candidates or staff, it may be chargeable for violating FEHA if the usage of the ADS creates disparate affect.
The laws warning that the usage of an ADS that, for instance, measures an applicant’s ability, dexterity, response time, and/or different talents or traits might discriminate towards people with sure incapacity or different protected traits. § 11016(c)(5). Equally, an ADS that, for instance, analyzes an applicant’s tone of voice, facial expressions, or different bodily traits or habits might discriminate towards staff or candidates based mostly on race, nationwide origin, gender, incapacity, or different protected traits. § 11016(d)(1). Accordingly, to keep away from illegal discrimination, employers might have to supply cheap lodging to an applicant or worker in step with the FEHA’s spiritual creed and incapacity protections. §§ 11016(c)(5) and (d)(1).
The Laws present that, to defend towards a discrimination declare based mostly on the usage of an ADS, employers can present that they carried out “anti-bias testing or related proactive efforts to keep away from illegal discrimination” previous to and after adopting an ADS. § 11009(f). The laws establish six related facets of such testing, together with the standard, efficacy, recency, and scope of such testing, in addition to the outcomes of the testing or different due diligence and the employer’s response to the outcomes (e.g., whether or not and the way the employer responded to the outcomes).
The laws have been additionally amended to require employers and lined entities to protect personnel and different employment information topic to the next necessities:
- Information have to be retained for a minimum of 4 years from the later of (a) the date the document was made, or (b) the date of the personnel motion — a rise from the earlier requirement to protect information for 2 years;
- Information topic to this requirement embrace choice standards, automated determination system information, functions, personnel information, membership information, employment referral information, and different information “created or acquired by the employer or different lined entity coping with any employment apply and affecting any employment advantage of any applicant or worker;” and
- “Automated-decision system information” consists of (a) any information utilized in or ensuing from the applying of an ADS, corresponding to information offered by or about particular person candidates or staff, or information reflecting employment determination or outcomes, and/or (b) any information used to develop or customise an ADS to be used by a selected employer or different lined entity.
The laws lengthen legal responsibility for ADS-driven discrimination to an employer’s “agent,” which is outlined as anybody “appearing on behalf of an employer, instantly or not directly, to train a operate historically exercised by the employer or another FEHA-regulated exercise,” corresponding to applicant recruitment, screening and hiring, promotion, or choices relating to pay, advantages, or go away, “together with when such actions and choices are performed in entire or partially via the usage of an automatic determination system.”
Employers that use ADS techniques in making employment choices and Al distributors whose merchandise are used within the employment area ought to take concrete steps to arrange for the implementation of the brand new Laws on October 1, together with:
- Figuring out the ADS techniques utilized in employment decision-making;
- Reviewing and amending document retention insurance policies to make sure information are retained for a minimum of 4 years;
- Performing anti-bias testing, establishing a plan outlining the frequency and nature of such testing, and documenting the testing course of/standards, outcomes, and steps taken to deal with outcomes, as applicable; and
- Updating anti-discrimination and cheap lodging insurance policies to deal with the usage of ADS techniques.
Arsen Kourinian and Ruth Zadikany are Companions and Remy N. Merritt is an Affiliate at Mayer Brown. This publish first appeared on the agency’s weblog.
The views, opinions and positions expressed inside all posts are these of the writer(s) alone and don’t signify these of the Program on Company Compliance and Enforcement (PCCE) or of the New York College College of Regulation. PCCE makes no representations as to the accuracy, completeness and validity or any statements made on this web site and won’t be liable any errors, omissions or representations. The copyright of this content material belongs to the writer(s) and any legal responsibility with reference to infringement of mental property rights stays with the writer(s).