California's No Robo Bosses Act and New Mexico's forthcoming frontier-AI framework mark two states moving on distinct fronts of AI regulation in the same week, filling a federal vacuum with approaches different enough to complicate compliance for any employer operating across state lines.
Why it matters
Governor Gavin Newsom signed SB 947 on Wednesday, banning California employers from relying solely on artificial intelligence to fire or discipline workers1,2,3. The same day, New Mexico Attorney General Raúl Torrez prepared to propose a state framework regulating frontier AI developers, requiring safety-risk disclosures, incident reporting, and independent audits4. ANALYSIS One law governs how employers use AI on their workforce; the other governs how developers build and release AI models. Together they illustrate a regulatory landscape fragmenting by both subject matter and geography.
The big picture
The California law prevents businesses from exclusively using "automated decision-making systems" in matters of employee discipline and termination. It also restricts employers' capacity to use AI as a "principal tool" in those decisions. Employers that rely primarily on AI output to make termination or disciplinary decisions must have a human reviewer corroborate those decisions using additional information. Affected employees must be provided with written notice that AI was primarily used, a description of the employee data the system consumed, and a human point of contact who can explain the decision.
Beyond the robo-boss ban, the package Newsom signed bans employers from using AI to predict a worker's emotional state through biometric data and requires written notices to workers if AI is responsible for mass layoffs. Newsom also signed a measure providing a state framework for independent evaluation and auditing of AI models.
New Mexico's proposed bill, set to be sponsored by Democratic state Rep. Linda Serrato, would require frontier AI developers to disclose safety risks, report serious safety incidents, and, for the largest companies, give advance notice of major training runs and submit to independent audits of their risk assessments. It would empower the attorney general to impose civil penalties for compliance violations, recover costs to the state for responding to safety incidents, and sue for damages on behalf of the state and its residents.
Between the lines
Newsom's signature itself tells a story of political recalibration. He vetoed an earlier version of the No Robo Bosses Act last October despite the legislation clearing both of California's legislative chambers with overwhelming majority support. The bill's author, State Senator Jerry McNerney, removed the pre-notification requirement when re-introducing the bill in February after Newsom's resistance. McNerney also stripped language that would have extended the Act's protections to gig workers. ANALYSIS The final law is narrower than the version Newsom rejected, suggesting the governor's reversal was enabled by concessions rather than a wholesale shift in posture.
Newsom was "sharply critical of Donald Trump for not passing comprehensive federal AI regulations". That framing positions state action as a response to federal inaction. ◆ Yet the emerging state-by-state approach is producing rules that vary in scope and target: California's law constrains employers deploying AI tools; New Mexico's bill would constrain the developers building frontier models. Illinois' Public Act 103-0804, which went into effect this past January, requires employers to notify workers when they use AI for specified employment purposes. New York became the first state to enact a ban on the construction of new AI data centers earlier this year. Each state is regulating a different link in the AI supply chain, and multi-state employers or developers face an accumulating set of obligations with no federal floor or ceiling to harmonize them.
The workplace pressure behind California's law is visible in the data. The OECD published survey results late last year finding that 90% of managers in the U.S. say their firms have adopted at least one tool to instruct, monitor, or evaluate workers. Meta is facing a lawsuit filed in July in which former employees allege AI-assisted systems were used to rank and select workers for layoffs. A May survey released by nonprofit United for Respect found that Walmart and Amazon workers are increasingly concerned that HR decisions are being automated.
U.S. Senators Ed Markey and Brian Schatz introduced federal legislation this past June with the same name as the California bill; Markey's bill would bar employers from relying on automated decision systems to make work-related decisions. ANALYSIS Federal legislation remains aspirational while state laws are now operational, widening the gap between where policy is debated and where it is enforced.
What's next
New Mexico's bill is set to be unveiled Thursday for the 2027 legislative session. In California, Newsom signed a broader executive order earlier this month addressing potential existential risks posed by AI models. Lorena Gonzalez, president of the California Federation of Labor Unions and lead sponsor of the No Robo Bosses Act, has framed the law as a baseline. ◆ With at least four states now operating distinct AI regulatory regimes and federal legislation stalled, the compliance map for employers and developers will grow more complex before any harmonization arrives.