State & Local GovernmentIssue 424 records
State Government
California launched a GenAI Skills Matching tool on CalCareers Oct 19, the first live service built on its AskCA engine; it suggests job classifications from an uploaded resume, stores nothing, and drew about 8,000 submissions in soft launch, with no accuracy or hiring-outcome data yet. Kentucky's May 2026 Personnel Memo 26-09, newly relevant, bars AI from application review, screening and selection while allowing HR drafting uses. Illinois rehired Tammy Roust as chief AI officer after her predecessor quietly left within about a month. Andalusia, Spain, began a ten-month rollout of 36,000 more Copilot-enabled licenses tracked by named indicators. One pattern: states are keeping AI on the applicant's side of hiring and out of screening decisions.
- Evidence records
- 4
- Cross-source patterns
- 1
- Evidence classes
- 2 independent reporting1 government evaluation1 standards or public-body guidance
- Outcomes
- 2 emerging2 cautionary
- Source freshness
- 3 new this fortnight1 older, newly relevant
- Research completed
- 2026-10-20
Choose a role to see its takeaway beside every record in the ledger.
Patterns across the evidence
Synthesis · Lighthouse Advisory interpretation
States are putting AI on the applicant's side of hiring while keeping it out of screening and selection
Interpretation. California's new CalCareers tool uses GenAI to suggest job classifications to applicants who upload a resume, without ranking them for agencies and without storing their data. Kentucky's Personnel Cabinet bars AI from reviewing applications, screening candidates or making selection decisions, while allowing drafting and other non-decision HR tasks. Both draw the same line: AI can help people find and prepare for state jobs, but decisions about candidates stay with people. Agencies adopting AI hiring tools should write that boundary down and check vendor features against it.
Operating questionFor each AI feature in your hiring process, is it helping applicants or deciding about them, and where is that boundary written and checked?
Supporting evidenceCalifornia Department of Human Resources (CalHR); California Office of Data and Innovation (ODI); Office of the GovernorKentucky Personnel Cabinet (Secretary Mary Elizabeth Bailey); Commonwealth Office of Technology
Evidence ledger
Full record · every source keeps its link and limitations
California launches a GenAI Skills Matching tool on CalCareers that suggests state job classifications from an uploaded resume; it is the first live public service built on the AskCA engine
On Oct 19 California announced a GenAI-assisted Skills Matching prototype on CalCareers, built by CalHR and ODI on the same technology as the AskCA prototype. Job seekers upload a resume and receive suggested state job classifications. The state says resume and conversation data are not stored and each session starts fresh. During the soft launch the tool drew about 21,500 clicks and nearly 8,000 resume or information submissions. CalHR calls it the first department deployment of an AskCA-powered tool in a live public service.
Why it matters, evidence and limitations
- Why it matters
- State hiring classifications are a common barrier for applicants, and AskCA (archived at its Sep 9 beta) has now moved from beta recruitment into a production-facing use. The tool advises applicants rather than screening them, which is a design choice other states can copy while avoiding the bias exposure of agency-side screening.
- Evidence and measured results
- Operator-reported usage only: about 21,500 clicks and nearly 8,000 submissions during the soft launch. No accuracy measure for classification suggestions, no applicant outcome data (applications, hires) and no evaluation method are given. 'Early testers called [it] especially helpful' is an anecdotal operator claim. The Accelerated Hiring Guide figures in the same release (500,000+ staff hours a year, 80-85% shorter screening) describe a process change, not the AI tool.
- Limitations and uncertainty
- Government press release; no independent evaluation. Usage counts are soft-launch engagement, not outcomes. The model, vendor and hosting are not named. The term 'prototype' signals the tool may change.
Kentucky Personnel Memo 26-09 bars AI from reviewing applications, screening candidates or making selection decisions in Executive Branch hiring, while allowing non-decision HR drafting uses
Kentucky's Personnel Cabinet directed, effective immediately on May 19, 2026, that AI tools are not approved for reviewing applications, screening candidates or making selection decisions in Executive Branch hiring, citing potential discriminatory bias against protected classes and equal employment opportunity law. HR staff may use AI for non-decision tasks: drafting job postings and position descriptions, template letters, summarizing non-confidential data and sample interview questions, under COT policy CIO-126. Entering applicant data, PII or other sensitive records into any AI tool is prohibited, and the rules will be added to statewide hiring training. Why now: California's Oct 19 launch of applicant-side GenAI job matching makes the line between helping applicants and screening them a live design question for other states.
Why it matters, evidence and limitations
- Why it matters
- It is a clear, citable state control that separates permitted HR drafting uses from prohibited selection uses, a template other HR departments and CIOs can adapt as AI hiring tools spread.
- Evidence and measured results
- Policy text only; no incident or bias finding is cited as the trigger. Permitted and prohibited uses and the data-entry ban are stated directly in the memo.
- Limitations and uncertainty
- Issued May 19, 2026, before this stream's recent runs; included for its relevance to the Oct 19 California launch. Enforcement method, monitoring and exceptions are not described. The memo applies to the Executive Branch.
Illinois rehires former CDO Tammy Roust as chief AI officer after the prior CAIO, tapped in June, quietly left after about a month; DoIT did not comment
StateScoop reported Oct 19 that Illinois DoIT rehired Tammy Roust, the state's chief data officer from November 2019 to October 2020, as chief AI officer this week. Her predecessor, Kader Sakkaria, tapped in June, appears to have left after about a month; comptroller records show his last pay date as July 24 and $9,000 in total pay. His LinkedIn shows a July move to Marsh. DoIT did not respond to StateScoop before publication. DoIT said on LinkedIn that Roust will guide the statewide strategy for safe, responsible and innovative AI use.
Why it matters, evidence and limitations
- Why it matters
- Illinois has active AI governance commitments (the archived EO 2026-07 AI Cabinet and an AI office the CIO called mature on Oct 2), yet its CAIO seat was effectively vacant from late July without public notice. Leadership continuity is a practical risk for state AI programs and for partners whose engagements depend on a named executive sponsor.
- Evidence and measured results
- Reported facts: hiring this week; predecessor's June appointment, July 24 last pay date and $9,000 total from comptroller records; Marsh move from LinkedIn. Roust's prior claim of a hospital reporting pipeline that cut staff time by 96% comes from her LinkedIn and is self-reported, not verified.
- Limitations and uncertainty
- Based on reporting with no DoIT comment on the departure; dates rely on comptroller and LinkedIn records. Roust's priorities are not yet stated. Whether the vacancy delayed any program is unknown.
Andalusia (Spain) begins rolling out 36,000 more Microsoft 365 and generative AI licenses including Copilot Chat to its general administration, tracked by adoption, training, productivity, security and service-quality indicators
On Oct 19 Andalusia's AI and digital minister told the ADA board that the regional government began deploying 36,000 new Microsoft 365 and generative AI licenses in October, at about 3,600 a month over ten months, adding to roughly 8,000 already active for a total of 44,500 users. Copilot Chat will support search, drafting, preliminary document analysis and content generation, with automation of document extraction, preliminary classification and case-file management to follow. The government says AI supports employees without replacing their judgment or responsibility, and lists governance controls on security, information protection, traceability and EU compliance, a basic training path with IAAP, a Digital Ambassadors network, and indicators for adoption, training, productivity, security and service quality.
Why it matters, evidence and limitations
- Why it matters
- A phased, ten-month enterprise copilot rollout to a state-sized workforce with named monitoring indicators is a useful comparison for U.S. states planning Copilot or Gemini expansions. Transferability is limited by EU AI Act and GDPR obligations and by Spain's administrative structure.
- Evidence and measured results
- Rollout scale, pace and planned indicators are government statements as reported. No baseline, targets, productivity results or costs are published.
- Limitations and uncertainty
- Spanish-language press report of a government note; the government note itself was not opened. Indicators lack definitions and targets. Licensing cost is not stated. Transferability to U.S. states is limited.
How to read this edition
Source findings, measured results and limitations come from the cited publications. Patterns, operating questions, role takeaways and implementation considerations are Lighthouse Advisory interpretation, stated as questions to validate locally rather than guaranteed outcomes. Vendor and operator claims are labeled as claims. Full research method.
- Independent reporting
- Independent reporting with attributable sources but without a formal evaluation design.
- Government evaluation
- A public body’s measured evaluation or documented pilot.
- Standards or public-body guidance
- Normative or advisory guidance from a standards body or public institution.