California's 2027 'No Robo Bosses' Law Leads a Wave of State AI Rules

California will ban sole reliance on automated systems in discipline and termination decisions from July 1, 2027, under the No Robo Bosses Act (SB 947). The law is the most concrete marker yet of a broader state turn: regulators are moving from broad AI risk-management principles to specific rules on disclosure, human review and layoff reporting.

SB 947 prohibits employers from relying solely on automated systems when disciplining or firing workers. Where an employer "primarily relies" on such a system, a person must corroborate its output against other evidence, such as personnel records or performance data. The employer must also give the affected employee written notice after the decision. The statute leaves "primarily relies" undefined, and that gap will test compliance efforts before the effective date arrives.

Colorado points in a similar direction. Its SB 26-189, effective Jan. 1, 2027, covers adverse AI-assisted decisions such as rejecting an applicant or offering lower pay. It requires a plain-language explanation of the technology's influence within 30 days. Affected individuals may seek correction of inaccurate personal data and meaningful human review where commercially reasonable.

What remains unsettled in Colorado?

Plenty. At an Oct. 26, 2026 hearing on the attorney general's proposed implementing rules, business groups argued the draft definition of "material influence" could trigger notice and human-review duties even when a person makes the final call. The Colorado Chamber of Commerce warned that notice and review for large numbers of unsuccessful applicants could burden high-volume hiring. Consumer advocates countered that the rules could limit when human review is available.

Litigation adds another layer. In X.AI LLC v. Weiser, the attorney general agreed to defer enforcement until 14 days after a ruling on a preliminary-injunction motion following final rules. Employers face a moving target on both dates and scope.

California's CCPA regulations, effective Jan. 1, 2027, grant related rights for certain significant employment decisions: opt-out of qualifying automated decision-making and access to information about how the technology was used. For some decisions, a business may offer a human appeal instead of the opt-out.

Does human involvement erase discrimination risk?

No. Since Oct. 1, 2026, Connecticut law expressly states that using automated employment technology is not a defense to a discrimination complaint. California's Fair Employment and Housing Act regulations address discriminatory automated decision-making, including anti-bias testing and when vendors act as an employer's agent. Illinois prohibits AI use that discriminates against protected classes.

Federal law still matters, though its posture has shifted. The EEOC's June 4, 2026 enforcement plan disavows disparate-impact litigation, but Title VII's disparate-impact provision remains intact and private claims continue. In Mobley v. Workday, Inc., an applicant alleges discrimination involving an AI screening system, and the case against the vendor remains pending. Human oversight does not close this exposure.

What must employers disclose?

The notice obligations are stacking up across states:

• Illinois requires notice to affected employees and applicants of covered AI use, though proposed regulations detailing the requirements were withdrawn.
• CCPA regulations require pre-use notice for covered automated decision-making.
• Colorado's SB 26-189 requires notice when individuals interact with covered decision-making technology.
• Connecticut, from Oct. 1, 2027, requires written notice identifying the AI tool, its purpose, the personal data used and a contact for inquiries.

Disclosure now reaches layoffs too. Connecticut requires employers filing federal WARN notices with its labor department to state whether layoffs relate to AI or another technological change. From Jan. 1, 2027, California's SB 951 requires qualifying Cal/WARN notices for reductions caused in whole or in substantial part by AI to identify affected occupations, automated job functions and the technology involved.

Colorado's earlier SB 24-205, which emphasized risk-management programs for algorithmic discrimination, was delayed and replaced before taking effect — a reminder that these frameworks are still being rewritten.

The bottom line

The new laws share one question: can a consequential decision be explained, reviewed and supported by evidence beyond an automated output? Employers should map where AI influences employment decisions, who reviews those outputs, and which notices or records each state requires. As AI becomes routine at work, the human role in decisions about people is set to matter more, not less.