Technology · Artificial intelligence · published 2026-10-08 · via ZDNET

California law reins in AI-only firing and discipline decisions

California Governor Gavin Newsom signed SB 947, a law aimed at limiting fully automated workplace discipline and termination decisions. The measure requires human review of automated outputs and says employers must give workers an explanation, including the data used, when such decisions are made. If the system's output cannot be confirmed or is judged unreliable, it must not be used for discipline or firing.

Expanded Detail

California's SB 947, signed by Governor Gavin Newsom, targets fully automated workplace discipline and termination. It requires a person to review such outputs and gives workers an explanation, including underlying data such as personnel records, evaluations, work product, peer reviews, and witness interviews. If the output cannot be confirmed or is deemed inaccurate, incomplete, or misleading, it cannot justify discipline or firing. The law takes effect July 1, 2027, and allows complaints to the state labor commissioner.

Workplace AI use is widespread: an April survey of 1,000 HR professionals found 73% use it in hiring, 52% in restructuring or role planning, and 51% to flag "risky" candidates. Only 26% said they avoid AI for layoff decisions. Federal rules remain less specific, while states vary; Colorado and Connecticut emphasize disclosure, and Illinois now requires third-party audits of frontier labs.

Count: First para: California's(1) SB(2) 947(3), signed(4) by(5) Governor(6) Gavin(7) Newsom(8), targets(9) fully(10) automated(11) workplace(12) discipline(13) and(14) termination(15). It(16) requires(17) a(18) person(19) to(20) review(21) such(22) outputs(23) and(24) gives(25) workers(26) an(27) explanation(28), including(29) underlying(30) data(31) such(32) as(33) personnel(34) records(35), evaluations(36), work(37) product(38), peer(39) reviews(40), and(41) witness(42) interviews(43). If(44) the(45) output(46) cannot(47) be(48) confirmed(49) or(50) is(51) deemed(52) i

Expanded detail and Context are AI-generated analysis; the linked article remains the authoritative source.

Read the full article at ZDNET →

This summary is Al-enhanced to contain extended analysis and broader social context. The original is {NAME); the linked article is the authoritative source. Original headline: “Regulators are trying to protect you from being fired by AI – here’s how.” Browse more stories.