1JUL

California binds AI hiring, EU defers

2 min read

11:00UTC

California advanced three AI-employment bills through committee as Brussels pushed its recruitment-AI compliance deadline back to December 2027.

← Back to Oracle names AI in its own annual report Jump to analysis ↓

EconomicAssessed

Key takeaway

California advances binding AI-hiring limits while the EU defers its equivalent workplace rules to December 2027.

California advanced three AI-employment bills through committee: SB 947, which bars firing or discipline by artificial intelligence (AI) alone, went to Assembly Appropriations on 1 July; AB 2545, creating a state study of AI's labour-market effects, cleared Senate Appropriations on 30 June; and AB 2656, requiring 45 days' notice to public-sector unions before any generative-AI rollout, was re-referred on 1 July⁠¹. The three continue the committee march this beat logged in late June.

Brussels moved the other way. The EU Digital Omnibus, given final Council adoption on 29 June, defers the high-risk compliance deadline for Annex III recruitment and employment tools, the CV-screening, candidate-ranking and termination systems, from 2 August 2026 to 2 December 2027⁠². The reprieve covers exactly the tools that decide who gets hired and fired.

In Washington state, the Washington Federation of State Employees returns to bargaining on 9 July, seeking contract language barring the University of Washington from using AI in hiring, promotion and dismissal decisions⁠³. Two regulatory models are now diverging in real time: US states and unions writing binding human oversight into law and contract, while the EU grants its own high-risk employment rules a sixteen-month reprieve.

Deep Analysis

In plain English

California is moving three separate bills through its legislature that would restrict how employers use AI to fire, discipline or monitor staff, including 45 days' notice to unions before rolling out generative AI. Meanwhile the EU has pushed back its own AI workplace rules by 16 months, from August 2026 to December 2027. The two moves point in opposite directions: California is tightening protection state by state while the EU, which was ahead on paper, has bought itself more time before its rules bite. Seattle's public-sector union is testing this gap directly, negotiating its own contract ban on AI hiring and firing decisions with the University of Washington.

Deep Analysis

Root Causes

The absence of a federal US AI-employment standard forces individual states to legislate against unverified attribution data, since no federal agency collects a national baseline comparable to what the EU's Digital Omnibus process is attempting.

The EU's legislative process, requiring Council and Parliament agreement across 27 states, structurally favours the lowest-common-denominator timeline, which the Digital Omnibus deferral exploited to push Annex III past its original date.

What could happen next?

Consequence
California's bills, if enacted, would give unionised US public-sector workers AI-notice rights ahead of any federal or private-sector floor.
Risk
The EU's second Annex III deferral weakens the credibility of future compliance deadlines, since firms now have precedent for delay if they lobby before a deadline lands.

Source Landscape

This story draws on neutral-leaning sources

Primary parallel: California's 2018 Consumer Privacy Act became the de facto US national privacy standard once no federal law followed, the same leadership pattern SB 947 and AB 2656 are attempting for AI-employment protection in the absence of federal legislation.

Counter-parallel: The EU's 2018 GDPR reached a single enforceable deadline across all 27 member states with no comparable multi-year deferral, unlike the AI Act's Annex III employment provisions, now delayed from August 2026 to December 2027.

Consensus view: Bruegel, the Brussels-based think tank, argues the Digital Omnibus deferral reflects a genuine measurement problem rather than regulatory retreat, since no EU member state has yet agreed a common method for attributing job loss to AI, the same definitional gap Challenger's US tracker struggles with.

Counter-view: Research ICT Africa, the Cape Town-based digital-policy institute, argues the deferral and California's patchwork bills both leave a protective vacuum that pushes AI-labour arbitrage toward jurisdictions with the weakest rules, since neither Sacramento's bills nor Brussels' delayed Annex III covers outsourced or offshore AI-affected work.

Key tension: Whether a binding EU-wide deadline or state-by-state US legislation is the faster route to enforceable protection, given both routes have now slipped their original timelines.

Sources:Gibson Dunn·Financial Conduct Authority·Bureau of Labor Statistics

Mentions:California →AB 2545 →AB 2656 →University of Michigan →Digital Omnibus →GDPR →Bruegel →Sacramento →Annex III →Brussels →Washington →Seattle →SB 947 →

First Reported In

Update #16 · AI layoffs fall, but the reversals begin

Gibson Dunn· 9 Jul 2026

Read original →

Causes and effects

Caused by

US states legislate as Washington stalls

The three bills continue the committee march this beat logged in late June.

Occurred 29 Jun 2026

Read story →

Brussels fixes AI Act job deadline

The EU Council's final adoption of the Digital Omnibus enabled the Annex III deferral now taking effect.

Occurred 29 Jun 2026

Read story →

This Event

California binds AI hiring, EU defers

US states are writing binding human oversight into law just as the EU delays its equivalent workplace rules by sixteen months.

Different Perspectives

Salesforce, Synopsys and TD Bank Group

Salesforce, Synopsys and TD Bank Group each filed quarterly disclosures in late August booking restructuring charges, or none at all, without naming AI as a cause. Their silence matters because Challenger's tracker shows AI as a stated reason fell to fourth place in August even as the year-to-date AI-cut total still leads at 116,175.

Singapore, South Korea, Taiwan and Indonesia

Singapore launched its Skills and Workforce Development Agency on 16 September, giving citizens six months of free premium AI tools, while South Korea ring-fenced its AI tax windfall in a new Future Response Fund. Taiwan kept funding its AI build past NT$190bn and Indonesia rewired vocational training around AI literacy, betting state-built skills beat a market-led adjustment.

ver.di, CGT Fonction Publique and CCOO

Germany's ver.di banked a 3.3% pay rise on 1 September and opened talks on a Tarifvertrag Transformation covering dismissal bans and reskilling, while France's CGT rejected Paris's AI negotiating timetable the same week. Spain's CCOO went further on 21 September, proposing to tax companies by the jobs they generate rather than wait for the next bargaining round.

BIS General Manager and Federal Reserve governors

The BIS's General Manager said on 10 September that AI displacement remains limited, even as the BIS's own survey found nearly 80% of firms plan to automate roles. Two Federal Reserve governors made the same point in July, arguing the labour-market data does not yet show a mass-firing event.

Bank of Canada, ONS and ECB

The Bank of Canada found the job-finding gap between AI-exposed and unexposed occupations widened from 2.2 to 13.9 percentage points since 2015-19, while separations barely moved. That framing, a hiring freeze rather than a firing wave, is echoed by the ECB's finding that euro-area AI use hit 52% of workers in 2026, concentrated among the university-educated.

Office for National Statistics

Deferred its Transformed Labour Force Survey beyond November 2027 and disclosed a May 2026 telephone-collection failure. The ONS carries no AI-attribution layer at all, so Britain sits outside this month's cohort of measuring states by its own admission.