California takes major step to regulate AI in the workplace

California Gov. Gavin Newsom signed a bill late last month that made the state the first in the nation to require human review for certain employment decisions made by automated systems. 

The “No Robo Bosses” Act, or SB 947, aims to reign in employers’ reliance on “any computational process,” including artificial intelligence, when determining whether to fire or discipline an employee, the bill reads. The legislation requires a human to “corroborate the decision” by reviewing the data used by or any other information relevant to the automated decision system’s output. Such information includes performance evaluations, peer reviews, personnel files and others. 

California’s “No Robo Bosses” Act is not the first state bill of its kind, “but it is, I think, the strongest that has been put forward in terms of regulating the use of artificial intelligence to make consequential decisions about employment,” Travis Hall, director for state engagement at the Center for Democracy and Technology, told Route Fifty. 

The bill “recenters an actual person in the decision-making … ensuring that there can be accountability for those decisions,” Hall said. The legislation helps remove mysticism surrounding the use of AI in the workplace and clarifies that someone can be held responsible for any incorrect or discriminatory employment decisions. 

It also grants employees in the state the right to be notified if a decision regarding their termination or discipline was assisted by an automated decision system and the right to request and receive “a meaningful, objective description of the employee’s own data” used by such tools. 

The bill also bars employers from using an automated decision system if a human reviewer cannot corroborate or verify a tool’s output. 

“No worker should ever be fired or disciplined by a robo boss. Artificial intelligence systems have the potential to increase productivity, but they’ve also been prone to errors, bias, and misjudgments,” California Sen. Jerry McNerney, a Democrat and sponsor of the bill, said in a statement last month. “The governor’s signing of SB 947 ensures human oversight of AI in the workplace. AI must remain a tool controlled by humans, not the other way around.”

California lawmakers introduced a similar version of the bill in 2025 that Newsom vetoed last October citing the legislation’s lack of specification. 

In a letter to the California State Senate explaining his veto decision, Newsom wrote that the “measure proposes overly broad restrictions on how employers may use [automated decision system] tools. For example, prohibiting an employer from using customer ratings as the primary input data for an ADS takes away a potentially valuable tool for rewarding high-performing employees.” 

Newsom also referred to “unfocused notification requirements on any business using even the most innocuous of tools” under the previous version of the bill, which required employers to notify staff before and after they used an automated decision system to inform a decision about a person’s employment, discipline or deactivation, according to the letter. 

States revisit AI regulation as concerns reemerge 

For many states like California, leading up to and throughout 2025, “there was a lot of energy and effort around the regulation of artificial intelligence … However, that momentum really stalled out,” Hall said, explaining that “there was concern that these [AI] regulations were going too far and would be onerous.”

As an example, Hall pointed to Colorado, which led such efforts after passing its landmark Artificial Intelligence Act in 2024 that looked to protect people from AI bias in employment, housing, healthcare, insurance and other decisions. 

The 2024 bill’s implementation faced delays during the rulemaking process and considerable industry opposition, including legal pushback earlier this year from the Justice Department and the company xAI, which is owned by Elon Musk. 

State-level AI regulation was also disrupted in 2025 when President Donald Trump tried to impose a 10-year moratorium on such bills, a move that the U.S. Senate eventually blocked in July. Trump signed an executive order in December in another bid to limit states’ efforts to regulate the technology on their own. 

This summer, Reps. Jay Obernolte and Lori Trahan also introduced the Great American AI Act, which if passed, would preempt certain kinds of states’ AI laws for three years. 

“I think that we're seeing a little bit of a turnaround in the [regulation] conversation where artificial intelligence is actually now being used, where new generations of artificial intelligence are being used to deploy, and the resistance to that is getting stronger,” Hall said.

Colorado lawmakers, for instance, passed SB 26-189 in May to repeal and replace the state’s original AI consumer protection bill with more stringent perimeters. SB 26-189, refers specifically to “automated decision-making technology” rather than “high-risk artificial intelligence systems” and loosens disclosure requirements for companies using such tools. 

A New Jersey Senate panel also advanced legislation aimed at regulating companies’ use of workplace tools used for personnel decisions earlier this month. The proposed legislation would also restrict employers’ use of monitoring tools in certain areas of the workplace, such as break rooms, and all areas outside of the workplace. 

New York lawmakers also considered a bill this year that, if passed, would have required organizations using AI to inform decisions about a person’s employment and other life situations to undergo regular audits. The bill aimed to reduce AI bias and discrimination in decision-making, particularly in the workplace, but it failed to pass into law before the legislative session ended. 

Alongside SB 947, Newsom also passed SB 813, which established a framework for third-party verification and assessments of AI systems, and AB 1405, which created a state registry and standards for AI auditors in the state. 

“AI is changing how we live, work, learn and connect, and Californians should have a real voice in what that future looks like,” Newsom said in a statement this week. “We’re listening directly to people across the state, gathering their input and making sure this technology is developed and governed in a way that expands opportunity, protects people and keeps Californians at the center of the decisions ahead.