{"resourceId":"unc-coates-local-ai-policy-recs-20260909","versions":[{"version":"external-f58067d265ddccdeff38959b1b11e5bf76b90a7c427314b0b11b735a731a0a80","resource":{"id":"unc-coates-local-ai-policy-recs-20260909","title":"UNC Coates’ Canons issues 14 recommendations for local-government AI use policies","organization":"UNC School of Government (Coates’ Canons)","sector":"Local-government law guidance and administrative policy","geography":"North Carolina / United States local governments","publishedAt":"September 9, 2026 Coates’ Canons blog post by Kristi Nickodem","publicationDate":"2026-09-09","eventDate":null,"sourceName":"Coates’ Canons (UNC School of Government)","sourceLabel":"University local-government law blog guidance on AI use policies","sourceUrl":"https://canons.sog.unc.edu/blog/2026/09/09/developing-and-updating-ai-use-policies-14-recommendations-for-local-governments/","evidenceClass":"standards-guidance","outcomeClass":"emerging","topics":["governance-procurement","data-security","operating-model","knowledge-work"],"finding":"On Sept. 9, 2026 UNC School of Government’s Coates’ Canons published Kristi Nickodem’s updated 14 recommendations for local-government AI use policies. The post argues many units still lack guardrails despite shadow AI: it cites a 2025 KPMG/University of Melbourne survey that 58% of employees intentionally use AI regularly and 70% of those use free public tools. Recommendations include: don’t copy-paste policies; define covered technology; require IT review for new software/subscriptions; set agentic-AI boundaries; prohibit improper confidential disclosures; require fact-checking and accountability for outcomes (noting LLM hallucinations); consider department-specific rules; remind staff of public-records duties; address transcription/notetaking/recording; state prohibited uses; calibrate transparency; involve the right people; review policies at least annually; and test whether employees can understand/apply the policy. Why-now for the 2026-09-23 edition: same-day Raleigh stop decision and Colorado Springs go-live make written fact-check, IT-approval, and prohibited-use rules immediately operational for peers.","sledRelevance":"Actionable local-gov policy checklist newly salient beside same-day production go-live and cautionary non-retention stories; fills the controls gap when cities buy voice AI or $3k LLMs.","evidence":"Coates’ Canons Sept. 9, 2026: 14 recommendations listed; shadow-AI rationale with 58%/70% survey cite; fact-checking/hallucinations; IT approval; confidential disclosures; agentic boundaries; annual review; employee understandability test.","architectureImplications":"Interpretation: Policies should distinguish chatbots, agentic tools, and transcription products so technical controls match risk.","governanceImplications":"Interpretation: Fact-checking, IT approval, prohibited uses, and annual review are minimum operating controls for both go-lives and stop decisions.","securityPrivacyImplications":"Interpretation: Ban confidential pastes into public tools; treat transcription/recording features as records and privacy events.","caveats":"Guidance blog, not an empirical evaluation of any city’s outcomes; survey statistics are secondary citations from 2025; older than last edition but included with explicit why-now beside 2026-09-23 deployments.","streamIds":["local-government"],"roles":{"sales":"Councils approving chatbots or $3k LLMs still need an AI use policy. Coates’ 14 recommendations support a discovery workshop on IT approval, fact-checking, confidential-data bans, and annual review—not a product endorsement. Bound engagements to policy drafting and employee comprehension tests before scale. Do not claim the post proves any vendor’s accuracy.","engineering":"Fit is policy-driven control requirements for staff and resident AI—not a model architecture. Translate recommendations into IT approval workflows, tool allowlists, logging, and human fact-check steps before outbound use. Prerequisites: inventory of shadow tools. Constraints: guidance only; no city outcome metrics. Security: block confidential disclosures into public LLMs; govern transcription tools.","delivery":"Draft/adapt a local policy, socialize with legal/IT/HR, train staff, and schedule annual review. Ownership: manager/attorney with IT. Skills: supervisors who enforce fact-checking. Governance checkpoints: council/manager adoption and 12-month review. Acceptance: employees can explain prohibited uses; IT gate exists for new AI subscriptions. Risks: copy-paste policies nobody follows and ignoring shadow AI."},"roleDetails":{"sales":{"customerProblem":"Employees already use public AI tools while many local governments still lack usable AI policies.","stakeholders":["City/county attorney","Manager/administrator","IT","HR","Department heads","Elected officials"],"discoveryQuestions":["Which AI tools are already in shadow use?","Who approves new AI software or subscriptions?","What fact-checking is required before external use?","When was the AI policy last reviewed?"],"valueHypothesis":"A tailored, understandable AI use policy reduces confidential disclosures and hallucinated communications if training and IT gates are real.","potentialEngagement":"Policy gap assessment and drafting workshop mapped to the 14 recommendations—no vendor accuracy claims.","unsupportedClaims":["Policy alone stops all misuse","Survey percentages prove local conditions","Any specific vendor is endorsed","One copied template fits every charter city"]},"engineering":{"fit":"Control framework requirements for generative and agentic tools used by local-government staff.","architectureIntegration":"IT service catalog allowlist → approval ticket → configured tenant/logging → human fact-check before external send.","prerequisites":["Tool inventory","Data classification","Records guidance","Training plan"],"constraints":["Guidance not empirical outcomes","Secondary survey citations","Must be localized to law","Older than last edition without why-now would be weak"],"security":"Prohibit confidential inputs to public tools; govern transcription/recording; review agentic write actions.","proofOfValue":["Share of AI tools IT-approved","Training completion","Fact-check spot audits","Annual policy review completed"]},"delivery":{"work":"Localize the 14 recommendations into a short policy, train, and set an annual review calendar.","dependencies":["Legal","IT","HR","Department champions"],"ownership":"Manager/attorney owns policy; IT owns technical gates.","skillsAdoption":"Tabletop with employees to test whether rules are understandable before publishing.","governanceCheckpoints":["Draft legal review","Adoption","90-day comprehension check","Annual update"],"acceptanceCriteria":["Covered technology defined","IT approval required for new AI software","Fact-checking required for external outputs","Annual review scheduled"],"risks":["Unreadable copy-paste policy","Shadow AI continues unchecked","Transcription creating unexpected records","Agentic tools acting without oversight"]}},"retrievedAt":"2026-09-24T03:22:00Z","enrichedAt":"2026-09-24T03:35:00Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Interpretation: Policies should be understandable to front-line staff, not only attorneys, or shadow AI will continue.","procurementImplications":"Interpretation: Route new AI subscriptions through existing IT acquisition gates; ban silent personal accounts for work data.","operatingModelImplications":"Interpretation: Pair go-lives and stop decisions with written fact-check and prohibited-use rules so accountability stays with humans.","sourceVerification":{"openedUrl":"https://canons.sog.unc.edu/blog/2026/09/09/developing-and-updating-ai-use-policies-14-recommendations-for-local-governments/","referenceExcerpt":"Require fact-checking and accountability for outcomes.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}