{"resourceId":"la-county-board-policy-6106-ai-continued-20261006","versions":[{"version":"external-46edb19f076b3c300df6ee92b5237c3080f583ab978984bb3d3a71ab3e88409a","resource":{"id":"la-county-board-policy-6106-ai-continued-20261006","title":"Los Angeles County AI Policy 6.106 bans fully automated high-impact decisions and unapproved AI tools, but Board vote is continued to Nov. 10","organization":"County of Los Angeles (Chief Executive Office / Office of the CIO; Board of Supervisors)","sector":"County AI governance policy","geography":"Los Angeles County, California, United States","publishedAt":"October 6, 2026 (board letter); supplemental agenda posted October 2, 2026","publicationDate":"2026-10-06","eventDate":"2026-10-06","sourceName":"County of Los Angeles Board of Supervisors (board letter and enclosed policy)","sourceLabel":"Government policy document","sourceUrl":"https://file.lacounty.gov/SDSInter/bos/supdocs/221879.pdf","evidenceClass":"standards-guidance","outcomeClass":"emerging","topics":["governance-procurement","data-security","accessibility-workforce"],"finding":"The county's CEO asked the Board to adopt Board Policy 6.106, which the Audit Committee approved Sept. 16. It sets eight principles, requires disclosure of AI-generated work product, and prohibits fully automated decisions that substantially affect people without meaningful human review, covert tracking or social scoring, manipulation, and the use of unapproved AI tools to upload non-public data, make official decisions, replace staff or produce authoritative work without human verification. It allows no exceptions. A supplemental agenda shows Supervisor Mitchell asked to continue the item to Nov. 10.","sledRelevance":"The largest U.S. county is codifying a prohibited-use list and named security and privacy roles, yet adoption has been deferred; this is a concrete template and a reminder that board timing is not guaranteed.","evidence":"Primary board letter and policy text, plus the supplemental agenda continuance entry; no implementation data.","architectureImplications":"The ban on unapproved AI sites and models for non-public data implies an approved-tool catalogue, data-classification enforcement and egress controls tied to existing Policy 6.104.","governanceImplications":"Interpretation: a no-exceptions policy with CIO, CISO and Chief Privacy Officer roles centralises accountability, but the continuance means departments operate under the earlier framework until at least Nov. 10.","securityPrivacyImplications":"Explicitly covers PII, PHI and Criminal Justice Information and assigns incident response to the CISO.","caveats":"Reasons for continuance not stated; printed sunset date 'September 31, 2030' is not a real date; Board minutes not inspected.","streamIds":["local-government"],"roles":{"sales":"Sales framing: Los Angeles County is close to a formal AI rulebook, but the vote has slipped to Nov. 10, so departments face uncertainty rather than a mandate. Useful conversations centre on readiness for the prohibited-use list: which tools are approved, how non-public data is kept out of others, and how AI-generated work is disclosed. Do not position any product as compliant with a policy that is not yet adopted, and avoid assuming the text will pass unchanged.","engineering":"Engineering framing: The policy translates into concrete controls: an approved AI tool catalogue, blocking or monitoring unapproved AI endpoints for county data, labelling of AI-generated outputs, and a human review step for any decision that substantially affects a person. Prerequisites are data classification under existing policy and identity-based access. Constraints include departmental IT autonomy and possible amendments before Nov. 10. Validate by testing whether current logs can show which AI services handled classified data.","delivery":"Delivery framing: Treat the continuance as preparation time. A delivery plan would map department AI uses against the prohibited list, define the human-review evidence each high-impact process must keep, and set up disclosure wording for AI-assisted documents. The Office of the CIO would own it with department CIOs. Hold any irreversible configuration until the Board acts, and track changes to the text, including correcting the invalid sunset date."},"roleDetails":{"sales":{"customerProblem":"Departments are using AI without an adopted countywide rulebook, and the proposed policy's prohibitions will require visible controls once approved.","stakeholders":["Board of Supervisors and deputies","Chief Executive Office","County CIO","CISO","Chief Privacy Officer","Department CIOs"],"discoveryQuestions":["Why was the item continued, and what changes are expected?","Which AI tools are approved today?","How will departments prove human review for high-impact decisions?","How is AI-generated work product disclosed now?"],"valueHypothesis":"Readiness work could shorten the time from adoption to compliance; no measured benefit is reported.","potentialEngagement":"A department-level gap assessment against the proposed prohibited-use list, delivered before the Nov. 10 hearing.","unsupportedClaims":["The policy has been adopted","Any tool guarantees compliance","The text will remain unchanged"]},"engineering":{"fit":"Fits policy-driven controls for approved tools, data classification and output disclosure; does not by itself define technical standards.","architectureIntegration":"Approved AI catalogue linked to identity and data classification, egress monitoring for unapproved AI services, output labelling and decision logs with reviewer identity.","prerequisites":["Data classification under Policy 6.104","Inventory of department AI uses","Central logging","Identity and access management"],"constraints":["Board approval pending","Decentralised department IT","Criminal justice and health data rules","Possible text amendments"],"security":"Keep PII, PHI and CJI out of unapproved services, enforce access controls, and route AI incidents through the CISO's incident process.","proofOfValue":["Sample audit of one department's AI tool use against the prohibited list","Test egress monitoring for unapproved AI domains","Check that high-impact decisions record a named human reviewer"]},"delivery":{"work":"Prepare departments to comply with Policy 6.106 if adopted: inventory, review evidence, disclosure templates and training.","dependencies":["Board adoption on or after Nov. 10","Department CIO participation","CISO and privacy office capacity","Final policy text"],"ownership":"Chief Executive Office / Office of the CIO, with department CIOs.","skillsAdoption":"Staff training on prohibited uses, disclosure of AI-generated work and when human verification is required.","governanceCheckpoints":["Board vote","Corrected sunset date and final text","First department compliance review"],"acceptanceCriteria":["Every department reports AI uses in a common inventory within a set period after adoption","Each high-impact AI-assisted process documents a human reviewer","Unapproved AI service use is measured and reported"],"risks":["Further delay or amendment","Departments treating 'no exceptions' as unworkable and bypassing it","Training lag"]}},"retrievedAt":"2026-10-06T22:14:13-05:00","enrichedAt":"2026-10-06T22:19:49.294-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"A workforce-empowerment principle favours retraining over displacement and bars using unapproved AI to supplant staff or contractors.","procurementImplications":"Policy applies to AI that is developed, procured or accessed through third-party platforms, so vendor terms must meet transparency and privacy requirements.","operatingModelImplications":"Department CIOs carry compliance duties under a central CIO-led AI Governance Framework.","sourceVerification":{"openedUrl":"https://file.lacounty.gov/SDSInter/bos/supdocs/221879.pdf","referenceExcerpt":"Fully automated decisions that do not require any meaningful human-in-the-loop but substantially impact individuals.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}