Feature
White Paper
The Future of Leadership Development: AI-Driven Manager Enablement
California's SB 947 is the state's most detailed attempt yet to limit artificial intelligence (AI) in the workplace. Its sponsors, State Senator Jerry McNerney and the labor federation backing it, prefer a less neutral name: the No Robo Bosses Act.
Introduced on Feb. 2, it revives SB 7, which cleared both chambers in 2025 before Governor Newsom vetoed it. The rewrite was intended to answer his objections, and it has had some effect: the Senate passed SB 947 by 29 to 9, and it now heads to the Assembly.
Two other bills are moving through the legislature with it: AB 1883 on surveillance and AB 1898 on disclosure. "Together, they create a two-layer regime," said Sarah Peterson Herr, a lawyer specializing in employment law and the senior legal editor at Brightmine. "You must disclose everything under AB 1898, and you cannot use certain tools at all under AB 1883."
Whether SB 947 survives is almost beside the point. Even if it fails, the framework these three bills propose — disclosure under AB 1898, prohibitions under AB 1883, human review under SB 947 — could easily resurface in future legislation. The result could be worker protections divided across state lines.
Table of Contents
- AI Assessments Are Already in the Workplace
- How AB 1898 and AB 1883 Interlock
- A Two-Tier Workplace?
- The Penalty Structure Across All 3 Bills
- What Comes Next
AI Assessments Are Already in the Workplace
The AB 1883 bill would bar AI used to assess emotional states, analyze gait, collect neural data or drive facial-recognition discipline.
The bill may be too little, too late. "All three of the technologies mentioned, emotion inference, gait analysis and using facial recognition to discipline, have been built into employee screening and monitoring software, even though employees will probably never know they're there," said Lacey Kaelani, CEO and co-founder of job search platform Metaintro. "Regulation is well behind the curve of what's already been implemented."
Peer-reviewed work supports the point. Workplaces are increasingly adopting and running emotion AI without workers' explicit knowledge, according to a 2023 study in the ACM CHI proceedings. IBM, Unilever, Microsoft and Softbank, among other employers, use emotional analytics for recruitment and monitoring. The latter research also notes the benefits are not well-established scientifically, which strengthens the case for regulating tools of unproven value.
When embedded in HR platforms, productivity software and workforce management systems, these tools make inferences workers cannot see, contest or confirm. The bill addresses this environment, one in which algorithmic management has become the norm. Under SB 947, employers could not rely solely on automated systems to discipline or dismiss and would have to apply human review and corroboration first.
How AB 1898 and AB 1883 Interlock
Alongside SB 947, the cleverest feature of the companion bills is not what they ban, but how they interlock. "An employer that deploys an emotion-recognition or gait-inference tool would first have to tell workers about it in detail under AB 1898, creating a paper trail that simultaneously proves a violation of AB 1883," Herr explained. "It becomes very difficult to use a banned tool covertly when the law separately requires you to disclose it."
Under SB 947, workers would gain the right to request a copy of any data used in a disciplinary or termination decision once a year, with written notice of what was used and who reviewed it. That access matters, because transparency works against employers, said Julian Gage of Engage Compliance. "A worker can simply point to the system the company deployed, and then discovery forces the employer to explain what it is inferring," he said.
None of it is new internationally, Gage added. The EU's GDPR and California's CCPA already restrict such inferences. "The real question for employers isn't the $500 figure. It's whether they can account for what their HR tools are inferring at all."
A Two-Tier Workplace?
While opponents warn that regulation will drive employers out of California, Kaelani treated this as posturing. The state is too large to leave, and its rules on privacy and the environment have historically set the national baseline, she said. The likelier response is selective non-deployment: running certain tools freely elsewhere while switching them off for staff in California, Herr predicted.
That points to a digital workplace fracturing along state lines, where the same software, under the same brand, manages a worker in Texas differently from one in California. For multinational companies already affected by GDPR, a patchwork management stack is familiar, but California would import it into the heart of the U.S. labor market.
The Penalty Structure Across All 3 Bills
What effect this might have comes down to enforcement, and to Gage's accounting problem: Most employers cannot say what their software infers about staff, because the capability is buried in tools bought for other reasons. AB 1898 forces those tools into the open; AB 1883 bans the most invasive ones; SB 947 requires human review before they drive a discipline or termination decision. Each carries its own penalty, but they're built on the same template.
On their own, the penalty figures look modest: $500 per violation under SB 947, and up to $500 per employee per violation under AB 1883 and AB 1898. But all three bills also let workers sue directly, with punitive damages and attorney's fees in play, which goes beyond the fine. And the per-violation specification means one platform used daily by thousands of workers, across multiple tools, can add up quickly, Herr said.
Whether the most novel bans can be policed is the harder question. A common objection holds that proving an employer used AI to infer a worker's mood is messy, and that such tools can simply be rebranded as ordinary "productivity analytics." Workplace software vendors frequently advertise these capabilities, so a worker can point to the product's marketing.
Federal law offers little cover either way. Nothing in Title VII, the ADA or the NLRA preempts a state from regulating workplace tools more tightly, and unlike federal law, the proposed regime applies whether or not discrimination can be shown, Herr said.
What Comes Next
Each bill is at a different stage. SB 947 has cleared the Senate and is now with the Assembly. AB 1883 and AB 1898 are moving in the opposite direction, working their way through the Senate after originating in the Assembly. Newsom's stated reason for vetoing SB 7 last year was that it swept too broadly, failing to separate high-risk tools from routine software, a concern business groups are pressing against the new package as well. The California Chamber of Commerce and the League of California Cities have already lined up against all three, Herr said.
The objection has thinned, however. California's Civil Rights Department and Privacy Protection Agency rules are in force, and they lean on audits that arrive long after a worker has already been disciplined or dismissed. The three bills try to move that intervention earlier: AB 1898 to before a tool is deployed, AB 1883 to the tools themselves and SB 947 to the moment a decision is made.
Webinar
On demand
Engaged, But Not Convinced: What's Eroding Beneath Your Strongest Engagement Numbers
Perceptyx EX Benchmark '26: leadership trust, change confidence and conviction gap truths at large enterprises.
Watch Now
Webinar
On demand
Retention Is a Communication Problem: New Data from 6,000+ Employees
This session gives leaders a clear diagnosis and practical path to treat retention as an infrastructure problem.
Watch Now
Webinar
On demand
Do Learning Programs Really Work? How to Turn Education Into Engagement In Healthcare
See how leaders are using learning programs to build trust with healthcare professionals and create measurable engagement.
Watch Now
Webinar
On demand
Fuel Your Front Line: Workforce Trends Shaping 2026
Get expert insight into the tools and tactics top frontline teams will need in 2026.
Watch Now
Webinar
On demand
Fix the Disconnect Between Projects and Profit
Your margins are hiding in disconnected data. Let's find them.
Watch Now
Webinar
On demand
AI in Customer Service: Faster Resolutions, Happier Customers
Don’t let rising demand burn out your team. See how to build a smarter, more resilient support org.
Watch Now
Any one of them may not pass. But the conditions they're reacting to don't need a bill to succeed in order to matter. The software making management decisions is already installed and already running, with or without a law requiring employers to disclose, restrict or review it. That is the condition all three bills are responding to, and it won't wait for either chamber to catch up.
Editor's Note: Regulation is playing a serious game of catch up with AI. Where else are we seeing that?
- Your Hiring Software May Already Violate EU Law — While the most consequential obligations of the EU AI Act may have been pushed off until December 2027, HR teams shouldn't see it as a reprieve.
- Why AI Hiring Discrimination Lawsuits Are About to Explode — AI is reshaping hiring — and the courtroom. Job seekers are suing over biased screening tools, and experts say a wave of lawsuits is just beginning.
- Your AI Notetaker May Already Be Breaking the Law — AI meeting notetakers are standard workplace tools. A new lawsuit suggests they shouldn't be treated that way.