California has signed new laws restricting some uses of AI in the workplace, including emotion-scoring tools and AI making decisions reserved for licensed health professionals. The rules matter to workers and employers alike: software already embedded in workplace systems may now face new limits.
Watch Desk analysis
What happened
A KQED report says Governor Gavin Newsom signed five labour-backed workplace AI bills at the end of September. Among them, AB 1883 restricts employers from using AI-powered workplace surveillance to read workers’ emotions or collect neural data. AB 1979 bars health facilities from letting AI alone perform work reserved for licensed professionals, and requires them to protect clinicians’ independent judgement when AI informs care.
The report describes a California nurse who said sentiment scores appeared in the metadata of recorded patient calls after her employer had paused a pilot. Kaiser told KQED it does not currently use AI tools to recognise, predict or infer nurses’ emotional states, and said it is reviewing AB 1883’s implications. The account points to a live dispute about what workplace systems do, not proof that the reported feature remains in use.
Why it matters
These laws reach beyond shiny new AI products. As KQED notes, employers may not know how much AI is embedded in software they already use, while the tools’ proprietary nature makes their use difficult to measure from outside. California’s rules put emotion inference and clinical responsibility among the questions organisations may need to answer.
The practical challenge is working out what a system actually does, who is accountable for its use, and whether a feature can be disabled. A policy on paper is a start; finding the switch is sometimes the more adventurous part.
Our read
This is a consequential development in AI policy because it targets specific workplace uses, not AI in the abstract. Employers should map what their systems infer and where AI informs clinical work. Workers, meanwhile, need clear explanations of what is being scored and why.
The laws’ reach will depend on their implementation and interpretation. But California has made one boundary explicit: AI should not take over work reserved for licensed professionals, and emotion-scoring at work is no longer just a question of management preference.
What to watch
- How California regulators interpret AB 1883’s restriction on emotion recognition.
- Whether employers disclose workplace AI features and how they are used.
- How health facilities demonstrate compliance with AB 1979.
- Whether other states adopt similar limits.
Discussion spark: Should employers be required to disclose every AI-driven score or inference made about workers, even when the feature is buried inside software they already use?
Sources and evidence
- Source update (28 October 2026, 00:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.