Lighthouse AdvisorySLED AI Adoption Intelligence

Health & Human ServicesIssue 033 records

Health & Human Services

Three sources, two from homelessness services. A UK government pilot added AI prompts to StreetLink public alerts and reports the share of rough sleepers found rising from 16% to 23%, now rolled out nationally (self-reported, no comparison group). Two days earlier the Centre for Homelessness Impact told councils to judge such tools on housing outcomes, not usage or hours saved. In Colorado, Gunnison County says an AI SNAP error-flag pilot is useful because staff review every flagged case. No new public health, public hospital, child welfare or aging evidence this edition.

Evidence records
3
Cross-source patterns
2
Evidence classes
1 government evaluation1 independent reporting1 standards or public-body guidance
Outcomes
1 effective1 emerging1 cautionary
Source freshness
3 new this fortnight
Research completed
2026-11-02

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

Patterns across the evidence

Synthesis · Lighthouse Advisory interpretation

2 patterns, each supported by at least two sources

  1. Homelessness AI is being judged on short-term service metrics while guidance asks for housing outcomes

    StreetLink reports better alert detail and a higher find rate after AI prompts, from a short self-reported pilot without a comparison group. Two days earlier, the Centre for Homelessness Impact told councils to judge AI tools on outcomes such as homelessness duties and sustained tenancies, use the strongest evaluation design possible, and publish null results.

    Operating questionFor an AI tool in homeless services, what outcome beyond usage or data quality will decide whether to scale it, and is there a comparison group?

    Supporting evidenceUK Ministry of Housing, Communities and Local Government (Homelessness and Rough Sleeping Directorate, Local AI team) with Home ConnectionsCentre for Homelessness Impact (funded by UK Cabinet Office Evaluation Task Force)

  2. AI is being placed upstream to improve inputs and target reviews, with people keeping the decision

    StreetLink uses AI to prompt the public for better referral details while outreach teams decide the response. Gunnison County uses AI to flag SNAP cases likely to contain errors, and staff review each one manually, citing local variable income. Neither source reports how accurate the flags or quality scores are.

    Operating questionHow will the agency measure whether AI flags and quality scores are accurate, and stop them from becoming de facto prioritisation?

    Supporting evidenceUK Ministry of Housing, Communities and Local Government (Homelessness and Rough Sleeping Directorate, Local AI team) with Home ConnectionsGunnison County Health and Human Services (Gunnison-Hinsdale Board of Human Services)

Evidence ledger

Full record · every source keeps its link and limitations

3 records
  1. Independent reportingEmergingNew this fortnight

    Rural Colorado county reports AI SNAP error-flag pilot is useful, with staff reviewing every flagged case

    At the Oct. 20 Gunnison-Hinsdale Board of Human Services meeting (reported Oct 28), Gunnison County, Colorado, staff said the county is participating in a SNAP accuracy pilot that uses AI to flag cases with a higher likelihood of error, which staff then review manually. Staff said it has been useful but stressed that human judgement matters when interpreting local, variable income. The same meeting described H.R.1 pressure: twice-yearly redeterminations and new work or volunteer requirements, with about 300 more redeterminations a year expected, amid turnover. Staff also said Colorado's county cohort model lacks shared technology, and the state's Office of Information Technology has not yet identified a vendor.

    Gunnison County Health and Human Services (Gunnison-Hinsdale Board of Human Services)Gunnison County, Colorado, United States2026-10-28

    Why it matters, evidence and limitations
    Why it matters
    With states now liable for more SNAP payment-error costs under H.R.1, county-administered states are using AI to target quality reviews. This is a frontline county view of how that works in practice.
    Evidence and measured results
    Staff statements at a public meeting as reported by a local paper. Qualitative only.
    Limitations and uncertainty
    Local newspaper summary of a board meeting; no metrics, pilot name, operator or vendor are given. Single small rural county. Whether the pilot is state-run is not stated.
    Local news report of board meeting Role takeaways & implementation
  2. Standards or public-body guidanceCautionaryNew this fortnight

    Centre for Homelessness Impact guide tells councils to judge AI homelessness tools on housing outcomes, not usage or hours saved

    The Centre for Homelessness Impact published (Oct 28) a guide, funded by the UK Cabinet Office Evaluation Task Force, on evaluating AI tools in homelessness prevention. It tells councils to ask first whether AI adds value over existing guidance, phone lines and webpages; to build in the strongest evaluation design they can, noting that digital tools make randomisation easier; to measure outcomes such as homelessness duties and sustained tenancies, not just usage, satisfaction or hours saved, and to expect a lag; to monitor safety incidents, escalations and accuracy; to count full costs including integration and upkeep; to keep a change log because tools change mid-evaluation; to treat chatbot logs as sensitive data with human escalation for high-risk cases, a Data Protection Impact Assessment and an Algorithmic Transparency Recording Standard record; and to share null and negative results. It uses an ongoing trial of one AI housing advice tool as a worked example; findings will be published separately.

    Centre for Homelessness Impact (funded by UK Cabinet Office Evaluation Task Force)United Kingdom2026-10-28

    Why it matters, evidence and limitations
    Why it matters
    U.S. human services and homelessness agencies piloting chatbots and AI tools mostly report usage, satisfaction and time saved. This guide gives a ready evaluation standard for asking whether housing outcomes actually improve. Transfer to U.S. programs is an interpretation.
    Evidence and measured results
    Recommendations from an evidence center; explicitly presents no findings.
    Limitations and uncertainty
    Guidance, not evidence of effectiveness. UK statutory framework (homelessness duties, DPIA, ATRS) differs from U.S. requirements. Only the publication landing page summary was inspected; the full guide is behind a link shortener and was not read.
    Practical guide (publication page) 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.

Government evaluation
A public body’s measured evaluation or documented pilot.
Independent reporting
Independent reporting with attributable sources but without a formal evaluation design.
Standards or public-body guidance
Normative or advisory guidance from a standards body or public institution.