California has taken a major step toward regulating artificial intelligence in the workplace, becoming the first state to prohibit employers from relying solely on an automated system to fire or discipline workers.
Gov. Gavin Newsom signed the No Robo Bosses Act, Senate Bill 947 (SB 947), on September 30, 2026. The law requires human oversight when an automated decision-making system plays a role in decisions that could cost an employee their job or lead to disciplinary action. California lawmakers describe it as the first law of its kind in the nation.
The law does not prohibit companies from using AI to manage their workforces. Instead, it draws a line around some of the most consequential employment decisions: a machine may assist with the decision, but it cannot be the sole decision-maker.
That distinction could become increasingly important as employers adopt software that monitors productivity, evaluates performance and analyzes employee behavior.
What does California’s No Robo Bosses Act do?
The No Robo Bosses Act targets what California law refers to as automated decision-making systems, or ADS.
Under SB 947, employers cannot rely solely on an ADS to make a decision to terminate or discipline a worker. When an automated system is used to assist with such a decision, a human must review and verify the outcome.
In practical terms, an employer cannot simply point to an AI-generated recommendation and treat it as the final answer.
A human decision-maker must remain involved.
The legislation is therefore less about banning workplace AI and more about establishing a human checkpoint before an automated recommendation can have serious consequences for an employee.
AI can still assist with workplace management
The law does not prevent employers from using AI-powered tools for workplace management.
Companies can continue to use automated systems for functions such as monitoring, analysis and productivity management, subject to other applicable laws and restrictions.
What changes is what happens when those systems become influential in a termination or disciplinary decision.
The basic framework is:
- AI or an automated system can assist with an employment decision.
- The system cannot be the sole basis for firing or disciplining a worker.
- A human must review and verify the decision when ADS is used to assist in termination or discipline.
- Employees must be informed when an automated system has been used in such decisions.
This approach reflects the law’s central premise: automated recommendations can inform a decision, but responsibility for a consequential employment action remains with a person.
Why is California regulating AI-driven firing decisions?
Employers are increasingly using software to evaluate employees, monitor workplace activity and measure productivity.
The concern is that automated systems can turn complicated workplace judgments into numerical scores or recommendations without necessarily capturing the circumstances behind an employee’s performance.
For example, an automated system could identify an employee as repeatedly missing performance targets. But the system may not understand why.
A manager reviewing the same situation could discover that:
- The employee was assigned unusually difficult work.
- The employee’s role changed during the evaluation period.
- A system used to measure productivity produced inaccurate data.
- The employee had documented performance issues that were already being addressed.
- Other managers or colleagues had information that the automated system did not consider.
The law’s human-review requirement is designed to ensure that those kinds of considerations are not automatically excluded from the decision-making process.
Sen. Jerry McNerney, who authored SB 947, said the measure is intended to ensure that AI remains a tool controlled by humans rather than replacing human judgment.
What information must employees receive?
Transparency is another major element of the legislation.
California’s official summary says employees must be informed when an automated decision-making system has been used in a termination or disciplinary decision.
The legislation also gives workers greater visibility into the information used by an automated system. Earlier legislative materials describing SB 947 said employees could request the data used by the system and the evidence supporting the resulting decision.
That matters because an employee challenging an AI-assisted decision needs more than a statement that “the algorithm determined” that disciplinary action was appropriate.
Without knowing what information was considered, an employee may have little practical way to identify an inaccurate record, incomplete information or potentially problematic automated assessment.
What are automated decision-making systems?
An automated decision-making system is not necessarily a humanoid “robot boss.”
The term can cover software that uses automated processes or AI to evaluate information and produce recommendations that affect people.
In a workplace, such systems can potentially be used for functions including:
- Employee performance evaluation
- Productivity monitoring
- Scheduling
- Workplace surveillance
- Risk or behavior assessment
- Attendance analysis
- Recommendations concerning discipline or termination
That is one reason the legislation is significant beyond generative AI chatbots.
A company does not need to use a system like ChatGPT to create an AI-related employment issue. An algorithm operating quietly in the background of an HR platform could have a far more direct effect on an employee’s job.
What does the law mean for employers?
For California employers, the legislation creates a new compliance consideration whenever automated systems become part of termination or disciplinary processes.
Companies using workplace AI may need to examine:
- What systems are being used?
Employers will need to understand which workplace technologies qualify as automated decision-making systems. - What decisions do those systems influence?
A tool used only for administrative tasks presents a different issue from one that influences whether someone keeps their job. - Who conducts the human review?
Employers will need an actual human decision-making process rather than treating a nominal review as a rubber stamp. - What information supports the decision?
Organizations should be able to identify the information considered when an automated system contributes to a consequential employment decision. - How are employees notified?
Employers must account for the law’s transparency requirements when ADS is used in discipline or termination.
The practical effect may extend beyond HR departments. Legal, compliance, information-security and technology teams could all have a role in determining whether workplace systems fall within the law and how their use should be documented.
Does the law ban “predictive” AI at work?
There is an important distinction between the final law and earlier versions of SB 947.
When the legislation was introduced in February 2026, it also proposed prohibiting automated systems from using workers’ personal information to predict future behavior.
That provision was part of the bill’s original framework, alongside the proposed human-review requirements.
The final legislation signed by Newsom is principally described by California officials in terms of preventing employers from relying solely on automated systems for firing or disciplining workers, requiring human oversight and verification, and informing workers when such systems are used.
That distinction is important when describing the law. The measure should not be characterized as a blanket prohibition on predictive workplace AI.
Why labor unions supported the legislation
The California Federation of Labor Unions, AFL-CIO, sponsored SB 947.
Labor advocates have argued that workers should not have to accept an automated employment decision without meaningful human accountability. They have also raised concerns about errors, bias and the difficulty employees may face when trying to challenge a decision generated or influenced by an algorithm.
Lorena Gonzalez, president of the California Federation of Labor Unions, has said workers and unions have pushed for stronger protections as workplace AI becomes more common.
The issue is not simply whether AI makes mistakes. Human managers make mistakes too.
The larger question is who is accountable when an automated system contributes to a consequential decision.
California’s answer is that a human must remain responsible for the decision.
Why this could become a bigger issue as workplace AI expands
AI is increasingly being marketed as a way to make workplaces more efficient.
Employers can use automated systems to process large amounts of information much faster than a manager could manually review it. That can be useful for identifying patterns and highlighting issues that might otherwise be missed.
But speed and scale can also magnify mistakes.
If a human manager makes one erroneous assessment, the impact may be limited to one employee. If an automated system contains a flawed assumption and is deployed across an entire organization, the same problem could potentially affect hundreds or thousands of workers.
That is particularly consequential when the output is connected to someone’s income.
California’s law therefore addresses a specific point in the AI decision-making chain: the moment when an automated recommendation could become a real-world employment action.
How will the law be enforced?
The law provides for enforcement by the California Labor Commissioner, the state attorney general and local prosecutors. The legislative summary states that SB 947 does not create a private right of action for employees.
That distinction matters.
It means the law creates regulatory enforcement mechanisms rather than simply giving every employee a new standalone lawsuit against an employer for an alleged violation.
Employers will therefore need to understand not only what their AI systems do, but also what records and procedures they maintain around their use.
What happens next for workplace AI?
California’s move could put additional pressure on employers and technology vendors to make workplace AI systems more transparent and easier for humans to review.
It could also influence the broader U.S. debate over whether AI should be allowed to make decisions affecting employment, housing, credit and other areas of people’s lives.
California has already adopted a number of AI-related workplace protections, and Newsom’s September 30 signing package included other measures addressing AI, automated technology and workplace surveillance.
The No Robo Bosses Act is narrower than a general workplace AI ban. Its significance lies in the decision it puts beyond the reach of an automated system acting alone.
A company can use technology to identify a problem. It can use an algorithm to flag an employee for review. It can use AI to process workplace information.
But when the consequence is discipline or termination, California law now requires a human to step back in.
TL;DR
- California Gov. Gavin Newsom signed SB 947, the No Robo Bosses Act, on September 30, 2026.
- The law bars employers from relying solely on automated decision-making systems to fire or discipline employees.
- When AI or another automated system assists with such a decision, a human must review and verify the outcome.
- Workers must be informed when automated systems are used in termination or disciplinary decisions.
- The law provides greater transparency around the information underlying automated employment decisions.
- The measure does not ban workplace AI generally.
- California officials describe SB 947 as the first law of its kind in the United States.
- The law could become a significant reference point as employers increasingly use AI for employee monitoring, evaluation and management.