Lighthouse AdvisorySLED AI Adoption Intelligence

Public Sector & Government · Issue 01 ·

Public Safety

Public-safety evidence baseline: a police reporting trial challenges assumed time savings, while two feedback trials show context-dependent changes in algorithm-defined speech scores. Ireland's newly operative High Court direction supplies document-specific verification controls. Corrections guidance emphasizes bounded evaluation, with no proven deployment benefit. Four canonical sources, two cross-source interpretations; historical research is explicitly dated and no archive items are recycled as new developments.

Evidence records
4
Cross-source patterns
2
Evidence classes
2 academic research2 standards or public-body guidance
Outcomes
2 emerging1 cautionary1 mixed
Source freshness
3 older, newly relevant1 undated
Research completed
2026-09-07

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. Validate the intended outcome, not a convenient proxy

    The reporting trial's null timing result and feedback trials' narrow scoring gains answer different questions. An agency should specify whether it seeks less work, better documentation or changed conduct, and evaluate that outcome independently before expansion.

    Operating questionWhat independent measure would establish improvement in the actual service objective?

    Supporting evidenceIan T. Adams and colleagues; Manchester Police DepartmentIan T. Adams, Kyle McLean and Geoffrey P. Alpert

  2. Human review needs source access, authority and correction routes

    HC 142's verification duties and the corrections brief's oversight recommendations support designing review as an operational workflow. Source access, accountable ownership and a usable correction process should be tested; the existence of guidance alone is not outcome evidence.

    Operating questionCan the responsible reviewer inspect the original evidence, stop an action and correct the record?

    Supporting evidenceHigh Court of IrelandUrban Institute; David Pitts and KiDeuk Kim

Full record · every source keeps its link and limitations

Evidence ledger

4 records
  1. Academic researchCautionaryNewly relevant · Oct 2024

    Police report-writing trial finds no significant time savings

    Access to Draft One did not significantly reduce report completion time in a preregistered Manchester trial.

    Ian T. Adams and colleagues; Manchester Police DepartmentManchester, New Hampshire, United StatesFirst published October 2, 2024; assigned to 2026 journal volume

    Why it matters, evidence and limitations
    Why it matters
    Baseline evidence newly added to this archive for police departments considering narrative copilots; this is not a newly conducted trial.
    Evidence and measured results
    85 officers were assigned to usual reporting (43) or AI access (42). Mixed-effects analysis covered 755 reports using RMS creation-to-submission timestamps; a 6,084-report annual analysis also found no significant effect.
    Limitations and uncertainty
    Single-agency intent-to-treat study: access did not mean use on every report. Elapsed timestamps are not active labor time. Report quality and downstream benefits were not established; a null result does not prove universal ineffectiveness.
    Peer-reviewed randomized field trial Role takeaways & implementation
  2. Academic researchMixedNewly relevant · Dec 2025

    AI feedback changes officer speech scores, with different results across two agencies

    AI-generated feedback changed algorithm-defined speech scores, with benefits differing by agency and feedback route.

    Ian T. Adams, Kyle McLean and Geoffrey P. AlpertAurora, Colorado, and Richland County, South Carolina, United StatesFirst published December 22, 2025; 2026 journal issue

    Why it matters, evidence and limitations
    Why it matters
    New-to-archive evidence for evaluating police coaching tools; not proof of reduced misconduct or improved public trust.
    Evidence and measured results
    Two six-month randomized trials enrolled 219 Aurora and 165 Richland officers, using no-feedback controls, self-assessment and supervisor-mediated arms. Analyses covered 124,443 and 65,172 videos. Aurora's two intervention arms reduced substandard scores; Richland's self-assessment arm increased high scores. Richland's supervisor-mediated result was not significant.
    Limitations and uncertainty
    The proprietary linguistic outcome lacked independent human validation. Scores are proxies, not established measures of procedural justice. Gaming and organizational context limit interpretation. Funding is attributed to the Laura and John Arnold Foundation.
    Peer-reviewed randomized trials Role takeaways & implementation
  3. Standards or public-body guidanceEmergingUndated source

    Irish High Court activates document-specific verification and AI disclosure rules

    HC 142 took effect September 1 for newly prepared civil court documents, requiring independent verification. Asking another AI to confirm accuracy is insufficient.

    High Court of IrelandIreland; transferable controls with U.S. jurisdictional limitsSigned July 29, 2026; public announcement August 31, 2026

    Why it matters, evidence and limitations
    Why it matters
    Current international court-control example; it is not binding on U.S. courts or a statement of Irish criminal-proceeding requirements.
    Evidence and measured results
    The direction preserves personal responsibility. General AI use does not automatically require disclosure, but court orders and document-specific duties apply. Witness evidence and expert reports have distinct declarations and requirements.
    Limitations and uncertainty
    No implementation outcome evaluation. Signature date is not confirmed web publication date. This resource addresses HC 142 only, not the separate Court of Appeal direction.
    HC 142 operative practice direction Role takeaways & implementation
  4. Standards or public-body guidanceEmergingNewly relevant · Feb 2026

    Corrections brief calls for bounded pilots and independent oversight

    The brief proposes limited corrections AI pilots with safeguards addressing bias, privacy and opaque systems.

    Urban Institute; David Pitts and KiDeuk KimUnited StatesLanding page dated February 4, 2026; PDF carries conflicting dates

    Why it matters, evidence and limitations
    Why it matters
    New-to-archive baseline guidance for state corrections agencies and county jails, rather than evidence of achieved rehabilitation or safety gains.
    Evidence and measured results
    The publication discusses administrative assistance and reentry applications and recommends oversight, data governance and independent evaluation. It provides no deployment comparison, measured benefit sample or causal impact estimate.
    Limitations and uncertainty
    Recommendations are normative. The landing page dates publication February 4, 2026, but PDF cover says February 2025 and copyright says December 2025. Exact PDF issue date remains unresolved.
    Policy brief, not an impact evaluation 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.

Academic research
Research produced through an academic institution or peer-reviewed venue.
Standards or public-body guidance
Normative or advisory guidance from a standards body or public institution.