What the law does
California just set a hard boundary for employer use of AI in discipline and terminations. SB 947, the No Robo Bosses Act, prohibits companies from relying solely on automated decision tools to fire or discipline employees and curbs the ability to make those systems the main driver in such calls.
When an automated system chiefly drives a discipline or firing call, a person must double‑check the result using additional sources like supervisor assessments, coworker feedback, and the employee's personnel records. In those cases, the employer must give the worker a written explanation stating that AI played the lead role, outline which employee data were analyzed, and provide a specific human representative to answer questions about the decision.
How it got here
McNerney, a Democrat, first introduced the idea in 2025 after a strong push from labor to set guardrails on AI in management decisions that hurt workers. The measure cleared both chambers with large majorities but hit a dead stop in October when Governor Newsom vetoed it. He warned at the time that while unregulated employer use of automated decision tools can harm workers, the bill then on his desk went too broad by requiring advance notices for even the most routine software.
McNerney brought the bill back in February and trimmed it. He removed the pre‑notification requirement that drew Newsom's objection and also cut language that would have covered gig workers, which had been blasted by Uber and Lyft. Those changes reduced, but did not erase, opposition from business and tech groups.
Among the strongest objections was from Robert Singleton, who serves as the Chamber of Progress' senior director overseeing policy and public affairs in California and the US West. In a letter urging a veto earlier this month, he argued the bill never defines when an employer "primarily relies" on an automated tool, leaving no clear line between software that merely informs a manager and software that becomes the main basis for a decision. He added that uncertainty about whether everyday tools count as regulated systems could chill the use of technology that improves consistency, flags safety risks, or helps managers make better choices.
Who backed it and how people are reacting
Organized labor was central.
The Electronic Frontier Foundation called the new law a "strong step toward giving workers the protections they need in workplaces that use automated decisionmaking systems."
When workplaces and rules change, staying steady with investments builds security, so get the free Always Be Buying E-Book today.
The politics are not cleanly partisan. This month, Steve Hilton, the Republican nominee for governor in California, backed the bill and criticized the California Chamber's position. "In some ways, I don't think SB 947 goes far enough," and he added that companies should be barred from using AI to determine whether a person is terminated, demoted, has their regular hours reduced, or is removed from an income program.
The wider AI moment and what it means for your money
Newsom has been active on AI policy lately. Earlier this month he signed a broad executive order focused on the potential existential risks posed by powerful models, saying the federal government was abdicating "its responsibility to protect Americans." He also approved a measure to set up a state process for independent evaluation and audits of AI models. All of that is unfolding as he weighs a possible 2028 presidential run and as AI policy stakes rise for Democrats and Republicans alike.
Public sentiment is shifting too. A July Gallup poll found 39% of Americans say AI does more harm than good, up from 31% in 2025. Pew Research reported in August that 71% think AI will "take people's jobs," about 7 percentage points higher than roughly two years earlier.
On the ground, the nonprofit United for Respect reported in May that Walmart and Amazon workers are increasingly worried HR decisions are being automated, and its push last summer for more AI disclosure at Walmart failed to pass. Meta also faces a July lawsuit from former employees who claim AI‑assisted tools sorted and chose workers for layoffs, and that the process disproportionately impacted individuals who had used medical or family leave. The company has denied those allegations.
The U.S. stands out for how widely algorithmic management has spread. An OECD survey published late last year reported that 90% of U.S. managers say their firms use one or more tools to direct, track, or evaluate workers.
California is now the first state with this level of oversight on workplace AI, going further than rules elsewhere. Illinois' Public Act 103‑0804, effective since January, requires notifying employees when AI is used for certain employment purposes but does not ban deploying agentic digital systems for those decisions. Other states are circling the issue.
Similar bills have been floated in New York, Louisiana, and New Jersey without action so far. And in Washington, U.S. Senators Ed Markey and Brian Schatz introduced a federal No Robo Bosses Act in June to stop employers from "relying on automated decision systems to make work‑related decisions," though it went nowhere in a Congress moving slowly on AI rules. California's step may be the nudge other states need.
For your portfolio, here is the takeaway: if a company builds, sells, or uses automated decision systems in HR, California just made documentation, disclosure, and a genuine human check part of the cost of doing business. That can affect compliance budgets, timelines for personnel actions, and legal exposure. As regulators and voters keep pressing on AI transparency and accountability, expect more companies to spell out how they combine software with human judgment, especially in states looking to follow California's lead.
Even as workplace rules and tech evolve, consistent investing pays off over time, grab the free Always Be Buying E-Book.
