{"resourceId":"porto-alegre-br-municipal-ai-law-14644-impact-assessment-20261022","versions":[{"version":"external-0701ed5f34db0762c8c322f60e62826eb3c89d6b4e8e146707cd6c8c8af6c321","resource":{"id":"porto-alegre-br-municipal-ai-law-14644-impact-assessment-20261022","title":"Porto Alegre enacts a municipal AI law requiring labelled generative content, human review of high-risk decisions and published impact assessments","organization":"Prefeitura de Porto Alegre (City of Porto Alegre)","sector":"Municipal AI governance legislation","geography":"Porto Alegre, Rio Grande do Sul, Brazil","publishedAt":"22 October 2026, 13:13 (city news release)","publicationDate":"2026-10-22","eventDate":"2026-10-22","sourceName":"Prefeitura de Porto Alegre","sourceLabel":"Official city news release on a signed law","sourceUrl":"https://www.prefeitura.poa.br/gp/noticias/prefeitura-sanciona-lei-que-estabelece-regras-para-uso-de-inteligencia-artificial","evidenceClass":"standards-guidance","outcomeClass":"emerging","topics":["governance-procurement","data-security","operating-model"],"finding":"Porto Alegre's mayor signed Law 14.644/2026, which governs AI use across the city administration. It requires generative-AI content to be clearly identified to the public, bars exclusively automated high-risk decisions with legal effects on citizens unless human review is possible, and requires a published Algorithmic Impact Assessment, reviewed at least every two years, before high-risk systems are deployed. A municipal data governance committee will oversee use, and the executive has 180 days to issue implementing rules.","sledRelevance":"Why now: signed 22 October and not previously archived. A city-level statute, not an internal policy, that codifies labelling, human review and impact assessment; useful comparison for U.S. councils weighing ordinances versus administrative policies, with limited transferability given Brazil's national legal framework.","evidence":"Official city release describing the signed law; statute text not opened.","architectureImplications":"Preference for 'responsive' AI in restricted, controlled environments points toward closed or grounded deployments over open internet models.","governanceImplications":"Moves AI rules into municipal law with a named oversight committee, published impact assessments and a 180-day regulatory deadline.","securityPrivacyImplications":"Impact assessments must address discrimination risk and affected groups; privacy provisions are not detailed in the release.","caveats":"Release-level summary; law text and future regulations not reviewed; Brazilian legal context limits direct U.S. transferability.","streamIds":["local-government"],"roles":{"sales":"Framing: AI rules are moving from policy memos into law. Porto Alegre's statute requires labelled generative content, human review of high-risk decisions and published impact assessments. Interpretation: for U.S. cities considering ordinances, this is a concrete comparator that supports discussions about assessment templates and labelling practice. It does not show that the law improves outcomes, and implementing regulations are still 180 days away, so it cannot be presented as a proven model.","engineering":"Framing: compliance has to be built into the system. Labelling generative outputs, routing high-risk decisions to human review and documenting systems for impact assessments are technical requirements. Interpretation: engineers should expect requests for output provenance tags, review queues and model documentation, and should favour grounded deployments in controlled environments as the law prefers. The release gives no technical standard, so specifics will depend on the coming regulations.","delivery":"Framing: departments now face a classification and documentation workload. Each system must be risk-classified, assessed before deployment and reassessed at least every two years. Interpretation: delivery teams should inventory existing AI uses, build an assessment process with the oversight committee, and plan for labelling across channels. Success should be judged by assessments completed and published on time, not by the law's existence."},"roleDetails":{"sales":{"customerProblem":"A city must turn a new AI law into working classification, labelling and assessment practice within 180 days.","stakeholders":["Transparency and Comptroller secretariat","Data governance committee","Department heads","City council","Legal counsel"],"discoveryQuestions":["Which current systems would count as high risk?","Who drafts and publishes impact assessments?","How will generative content be labelled in each channel?","What do the regulations need to specify?"],"valueHypothesis":"Bounded: a structured assessment and inventory process could help meet statutory deadlines; no outcome data exist yet.","potentialEngagement":"AI inventory and impact-assessment template development aligned to the statute.","unsupportedClaims":["That the law reduces harm","That current systems comply","Direct applicability to U.S. jurisdictions"]},"engineering":{"fit":"Municipal systems that make or support decisions with legal effects on residents, and any generative content published to the public.","architectureIntegration":"Output labelling, human review queues and documentation hooks added to existing AI services.","prerequisites":["System inventory","Risk classification criteria","Model and data documentation"],"constraints":["Regulations pending for 180 days","Brazilian legal framework","Preference for controlled environments"],"security":"Restrict high-risk systems to controlled environments and document data flows for impact assessments.","proofOfValue":["Label generative output in one public channel and audit coverage","Run one high-risk system through a draft assessment","Test the human-review path end to end"]},"delivery":{"work":"Stand up the inventory, classification, assessment and labelling processes the law requires.","dependencies":["Implementing regulations","Oversight committee capacity","Department cooperation"],"ownership":"Secretaria Municipal de Transparência e Controladoria with the data governance committee.","skillsAdoption":"Staff need training on risk classification, assessment writing and human-review duties.","governanceCheckpoints":["Regulations issued within 180 days","Assessments published before high-risk deployments","Biennial reassessment"],"acceptanceCriteria":["Proposed: complete inventory of AI systems in use","Proposed: every high-risk system has a published assessment before go-live","Proposed: labelling present on all public generative content"],"risks":["Assessments become paperwork","Regulatory delay","Under-resourced oversight committee"]}},"retrievedAt":"2026-10-25T22:17:48-05:00","enrichedAt":"2026-10-25T22:24:37-05:00","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Human review requirement for high-risk decisions keeps staff in the loop for consequential cases; staffing for assessments not addressed.","procurementImplications":"Vendors supplying high-risk systems would need to support pre-deployment impact assessment and human-review workflows.","operatingModelImplications":"Departments must classify systems by risk, label generative outputs and publish assessments before go-live.","sourceVerification":{"openedUrl":"https://www.prefeitura.poa.br/gp/noticias/prefeitura-sanciona-lei-que-estabelece-regras-para-uso-de-inteligencia-artificial","referenceExcerpt":"A lei também determina que conteúdos produzidos por IA generativa sejam identificados de forma clara ao público.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}