{"resourceId":"maryland-moore-ai-framework-frontier-audits-20260922","versions":[{"version":"external-d1f4999527589e4c499418c7f1373f9a6c73a9017d2206baa0cb59741baea98d","resource":{"id":"maryland-moore-ai-framework-frontier-audits-20260922","title":"Maryland Moore AI framework seeks frontier audits, 72-hour incident reporting, and worker voice","organization":"Office of Governor Wes Moore (Maryland)","sector":"State executive AI policy / workforce and consumer protection agenda","geography":"Maryland, United States","publishedAt":"2026-09-22","publicationDate":"2026-09-22","eventDate":"2026-09-22","sourceName":"Office of Governor Wes Moore — Governor Moore Outlines AI Framework to Protect Marylanders","sourceLabel":"Official state governor press release announcing AI principles and planned actions","sourceUrl":"https://governor.maryland.gov/news/press-releases/governor-moore-outlines-ai-framework-protect-marylanders","evidenceClass":"government-evaluation","outcomeClass":"emerging","topics":["governance-procurement","accessibility-workforce","data-security","operating-model"],"finding":"Why now: on 2026-09-22 Governor Moore announced an AI framework organized around protecting people, centering workers, and keeping kids safe. Under protecting people, the AI Subcabinet will develop legislative recommendations for regulating frontier AI model companies—including safety frameworks and testing, worker protections, independent third-party audits and evaluations, whistleblower protections, public reporting, and 72-hour incident reporting. Additional actions span Right of Publicity, civil-rights guidance for AI in housing/lending/employment, algorithmic tenant-screening limits on DHCD-financed properties, anti-rent-collusion legislation, gambling targeting limits, a standing Labor-Management Committee on AI for state workers, and stronger protections related to unsafe chatbots for youth.","sledRelevance":"State-level frontier-developer oversight agenda that mirrors elements other states are operationalizing (audits, incident reporting) while adding explicit state-workforce labor engagement—useful for multi-state policy and shared-services planning with clear 'not yet law' limits.","evidence":"Official press release dated September 22, 2026 (updated same day). Builds on 2024 EO, AI Governance Act of 2024, and Responsible AI Policy. No quantitative agency AI outcome metrics in the release.","architectureImplications":"If recommendations become law, frontier vendors serving Maryland may need audit/eval evidence and incident-reporting pipelines; state DoIT/DBM guidance for workplace AI may affect tool configuration and logging for state workers.","governanceImplications":"Creates a principles-to-legislation pipeline via AI Subcabinet plus cross-agency civil-rights and labor structures; distinguishes announced intent from enacted duties.","securityPrivacyImplications":"Emphasizes rights, privacy, deepfake/likeness protections, and incident reporting for frontier companies; workplace AI discrimination guidance implicates HR/systems data use.","caveats":"Most items are planned legislative recommendations or agency guidance tasks, not enacted requirements as of the release. Housing/gambling/chatbot provisions need future bill text. No measured results.","streamIds":["state-government"],"roles":{"sales":"Sales framing: Maryland is packaging frontier-company audits, 72-hour incident reporting, and state-worker labor voice as an executive agenda—not a finished statute. Sell policy-readiness and inventory work that anticipates those recommendations, especially for vendors already in state contracts. Keep claims bounded to 'prepare for likely requirements' and avoid implying the audits are already mandatory. A short discovery engagement on current frontier usage and incident processes fits better than a large build.","engineering":"Engineering framing: map Maryland’s frontier recommendations to concrete telemetry—model inventory, eval artifacts, whistleblower and incident channels, and workplace-AI logging under DoIT/DBM guidance. Keep companion-chatbot youth protections architecturally separate from enterprise productivity tools so controls do not blanket every Copilot seat. Build a gap matrix against independent-audit and 72-hour reporting concepts before any new high-risk deployment. Proof-of-value is that matrix plus a tabletop incident drill, not a chatbot launch demo. Treat the release as design input until statutes exist.","delivery":"Delivery framing: help DoIT/DBM and the AI Subcabinet sequence workplace-AI guidance and a legislative-recommendation packet without pretending the agenda is already binding law. Acceptance criteria should label each item as guidance versus bill language and name accountable owners, including the Labor-Management Committee pathway for state workers. Charter a youth-chatbot protection workstream separately from enterprise tools. Primary risks are over-implementing un-enacted rules, under-engaging unions, or letting housing/gambling algorithm workstreams starve IT capacity."},"roleDetails":{"sales":{"customerProblem":"States need a coherent frontier-AI oversight story spanning audits, incidents, workers, and kids without overstating what is already binding.","stakeholders":["Governor’s AI Subcabinet","DoIT / DBM","MCCR / labor agencies","Procurement","Union representatives"],"discoveryQuestions":["Which frontier models are in production today?","What incident clock exists now?","How will Labor-Management Committee review workplace AI?","Which chatbot youth rules already bind vendors?"],"valueHypothesis":"A bounded readiness assessment can align vendor and agency practices to Maryland’s announced frontier-audit and incident-reporting direction before bills land.","potentialEngagement":"4–6 week AI Subcabinet briefing pack and agency inventory against the framework’s frontier items.","unsupportedClaims":["Independent audits already legally required in Maryland","72-hour reporting already enforceable statewide","Guaranteed youth-chatbot compliance without bill text"]},"engineering":{"fit":"Fit for governance/telemetry design anticipating frontier audits and incident reporting; unfit as authority to deploy new high-risk systems.","architectureIntegration":"Inventory frontier vs non-frontier tools; define incident ticket schema; segregate workplace AI configs under DoIT guidance.","prerequisites":["Model/use-case inventory","Incident response runbooks","Labor consultation channel","Civil-rights guidance hooks"],"constraints":["Recommendations not yet statute","Maryland-specific civil-rights and housing contexts","Federal preemption uncertainty"],"security":"Incident and whistleblower channels must protect reporters and sensitive model-eval data.","proofOfValue":["Gap matrix vs audit/incident recommendations","Tabletop 72-hour incident drill","Workplace AI guidance pilot in one agency"]},"delivery":{"work":"Facilitate Subcabinet recommendation drafting support, worker-committee process design, and agency guidance rollouts.","dependencies":["AI Subcabinet","DoIT/DBM","MCCR/Labor","Legislative sponsors if bills advance"],"ownership":"Governor’s office / AI Subcabinet with agency implementers for guidance.","skillsAdoption":"Agency counsel and security staff need frontier-eval literacy; HR needs workplace-AI discrimination guidance fluency.","governanceCheckpoints":["Principles published","Recommendation draft","Labor-Management standing agenda","Bill introduction"],"acceptanceCriteria":["Written frontier recommendation outline including audits and 72-hour reporting","Named owners for workplace AI guidance","Youth chatbot protection workstream chartered"],"risks":["Agenda outruns legislative capacity","Confusing guidance with binding law","Under-engaging unions on workplace AI"]}},"retrievedAt":"2026-09-23T22:25:26-05:00","enrichedAt":"2026-09-23T22:33:54-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Framework centers worker voice, Labor-Management Committee, reskilling, and school-to-career pathways; accessibility for residents is implied via rights protections rather than WCAG metrics.","procurementImplications":"Frontier legislative recommendations (audits, incident reporting, whistleblower protections) would reshape vendor diligence if enacted; current release does not itself amend procurement manuals.","operatingModelImplications":"Creates a principles-led operating cadence across Subcabinet, civil-rights agencies, DoIT/DBM, and labor—parallel tracks for people, workers, and kids rather than a single IT project.","sourceVerification":{"openedUrl":"https://governor.maryland.gov/news/press-releases/governor-moore-outlines-ai-framework-protect-marylanders","referenceExcerpt":"independent third-party audits and evaluations, whistleblower protections, public reporting, and 72-hour incident reporting","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}