Lighthouse AdvisorySLED AI Adoption Intelligence

Public Sector & Government · Issue 05 ·

State Government

Three newly archived sources examine California's shared AI platform, North Carolina's forthcoming oversight playbook and independent scrutiny of state AI contract provisions. Platform capability claims and planned controls are not measured service gains. Historical dates and research limitations are explicit. Two cross-source interpretations address model-change evidence and operational review capacity. No fresh measured September service improvement was established; an international audit could not be accessed.

Evidence records
3
Cross-source patterns
2
Evidence classes
2 standards or public-body guidance1 independent research
Outcomes
2 emerging1 cautionary
Source freshness
1 undated1 new this fortnight1 older, newly relevant
Research completed
2026-09-11

Choose a role to see its takeaway beside every record in the ledger.

Synthesis · Lighthouse Advisory interpretation

Patterns across the evidence

2 patterns, each supported by at least two sources

  1. Shared model access needs evidence for each material change

    Poppy's interchangeable models make evaluation continuity relevant; the FAS memo's contract scrutiny raises the complementary question of whether agencies can obtain the evidence required. Neither source establishes a failure in Poppy's contracts. Test technical portability and evidence access together.

    Operating questionCan the agency retest, inspect and exit a workflow after its model or provider changes?

    Supporting evidenceCalifornia Department of TechnologyFederation of American Scientists; Jae Yeon Kim and Aniket Kesari

  2. Approval needs an operating review capacity

    North Carolina's intended inventory and monitoring lifecycle and FAS's proposed renewal disclosures both create continuing review work. Assign an owner and a decision process for that evidence. Neither source demonstrates that these controls already improve outcomes.

    Operating questionWho reconciles deployment changes, evaluates new evidence and decides whether continued use remains acceptable?

    Supporting evidenceNorth Carolina Department of Information TechnologyFederation of American Scientists; Jae Yeon Kim and Aniket Kesari

Full record · every source keeps its link and limitations

Evidence ledger

3 records
  1. Standards or public-body guidanceEmergingUndated source

    Poppy describes shared multi-model access without measured productivity evidence

    CDT describes a multi-model employee assistant with reusable workflows and API integration. Its security and productivity statements are operator claims, not independent test results.

    California Department of TechnologyCalifornia, United StatesUndated page; inspected September 11, 2026 UTC

    Why it matters, evidence and limitations
    Why it matters
    Direct state shared-service example, newly added to complement earlier California procurement coverage; current rollout completion is unverified.
    Evidence and measured results
    The page reports pilot input from 2,800+ employees across 67 departments. It dates the pilot start to September 29, 2025 and still describes July 2026 rollout as planned. Testimonials provide no measured baseline, quality-adjusted savings, sampling method or controlled evaluation.
    Limitations and uncertainty
    Undated and partly forward-looking. No independent security validation or measured service benefit. Event date refers only to pilot launch.
    Official platform guidance and operator claims Role takeaways & implementation
  2. Standards or public-body guidanceEmergingNew this fortnight

    North Carolina prepares lifecycle AI oversight with inventories and continuing monitoring

    NCDIT announces preparation for an AI Governance Playbook spanning assessment, approval, inventory, mitigation and monitoring; it does not report implementation results.

    North Carolina Department of Information TechnologyNorth Carolina, United StatesSeptember 2, 2026

    Why it matters, evidence and limitations
    Why it matters
    Direct state-agency preparation signal. The September 2 notice is newly archived, not a new September 10 launch.
    Evidence and measured results
    The notice specifies quarterly inventory submissions and planned publication of high-risk uses, alongside risk and privacy assessments. It provides no completed assessment sample, compliance rate, baseline or measured efficiency result.
    Limitations and uncertainty
    Prelaunch notice, not the complete playbook or proof that agencies comply. Exact launch date is unknown; no independent outcome evaluation.
    Official forthcoming governance guidance Role takeaways & implementation
  3. Independent researchCautionaryNewly relevant · Jun 2026

    Contract analysis raises accountability questions, with methodological and editorial limits

    The authors report limited AI-specific accountability language in sampled provisions and advocate risk-based purchasing and updated vendor disclosures. Recommendations are not evaluated interventions.

    Federation of American Scientists; Jae Yeon Kim and Aniket KesariUnited States; contract analysis covers California, Utah and FloridaJune 8, 2026

    Why it matters, evidence and limitations
    Why it matters
    Historical purchasing scrutiny newly added to qualify shared-service expansion and oversight preparation.
    Evidence and measured results
    The methods paragraph identifies 3,771 provisions across 215 contracts in three states and reports 77% as boilerplate. This is a provision-level statistic, not a percentage of unsafe systems or ineffective contracts.
    Limitations and uncertainty
    Ongoing research without a reproducible coding protocol or inter-rater reliability on the page. Broader sample wording is inconsistent; unrelated historical and cost statements also contain apparent errors and were excluded. No causal test of contract reform.
    Policy memo reporting ongoing independent research Role takeaways & implementation

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.

Standards or public-body guidance
Normative or advisory guidance from a standards body or public institution.
Independent research
Research conducted outside the implementing organization.