{"resourceId":"ccj-corrections-ai-classification-case-202609","versions":[{"version":"external-726c08f1687b5690843bd840801cfb341630be47db14496f7f5d75e5b3cd1c04","resource":{"id":"ccj-corrections-ai-classification-case-202609","title":"CCJ corrections case study ties AI classification to Title VI testing, contestability and annual reassessment","organization":"Council on Criminal Justice Task Force on Artificial Intelligence","sector":"Corrections / custody classification policy","geography":"United States state corrections (hypothetical DOC)","publishedAt":"September 2026","publicationDate":null,"eventDate":null,"sourceName":"CCJ","sourceLabel":"Council on Criminal Justice case study","sourceUrl":"https://counciloncj.org/assessing-ai-in-action-a-case-study-on-the-use-of-ai-powered-classification-tools-in-state-corrections","evidenceClass":"standards-guidance","outcomeClass":"cautionary","topics":["governance-procurement","data-security","accessibility-workforce","operating-model"],"finding":"September 2026 CCJ case study applies the User Decision Framework to AI-powered inmate classification supporting custody, housing, programming, and reclassification for a hypothetical 15k–25k person state DOC. It finds substantial risk and substantial opportunity, requiring Careful Implementation: Title VI-grounded disparate-impact testing as a contract term, case-specific decision rationales on contest, documented justification both when following and overriding scores, Community Advisory Committee with directly affected representation, independent validation, and annual reassessment. Prohibited-use screening passes only if disparate-impact testing and substantive contest can be negotiated; otherwise treat as prohibited until mitigated. Compares AI to existing arbitrary/under-explained baselines rather than ideal human judgment. Not a measured deployment evaluation.","sledRelevance":"Actionable for state DOC and local jail classification modernization: defines contract redlines (disaggregated accuracy, data portability, audit rights) and operational controls (override logging by demographic group). Interpretation; federal equal-protection intent doctrine alone is called inadequate—Title VI and state rules matter.","evidence":"Opened original HTTPS HTML. Screening table marks contest and disparate-impact as YES with mitigation required; risk section cites liberty impacts and documented disparities in existing tools (COMPAS/PATTERN references).","architectureImplications":"Integrate with jail/OMS, health, and programming systems carefully; demand case-specific rationales for higher-complexity/ML components; preserve human-in-the-loop overrides; separate classification modules from vendor surveillance/predictive-violence add-ons for distinct scrutiny.","governanceImplications":"Name a DOC official for classification outcomes and disparity monitoring; publish outcomes by demographic group; stand up compensated Community Advisory Committee with real recommend authority; freeze deployment on widening disparities.","securityPrivacyImplications":"CJIS compliance, minimum-necessary collection, no cross-system repurposing without new assessment, agency data ownership/portability against lock-in, no vendor training on agency data without consent.","caveats":"Hypothetical DOC scenario; evidence of opportunity draws on structured classification literature generally, not a specific AI product RCT; proprietary tools vary widely in transparency.","streamIds":["public-safety"],"roles":{"sales":"Sales framing: Position this as a corrections classification governance engagement, not a housing-algorithm sale. The customer problem is inconsistent, poorly audited placement decisions with liberty stakes—not missing AI. Bound the ask to whether Title VI testing, contestability, and data-portability terms are obtainable before any platform demo. Value hypothesis is reduced over-classification risk and auditable decisions versus today's opaque baselines—not guaranteed recidivism drops. Unsupported: claiming AI is fairer without disaggregated validation.","engineering":"Engineering framing: Fit only where placement officers remain in the loop and the vendor exposes factors and weights enough for individual contests. Architecture must log follow-and-override justifications, support demographic monitoring, and avoid auto-accept defaults. Prerequisites include data-mapping quality across OMS, EHR, and programming feeds plus independent validation rights at renewal. Constraints include trade-secret opacity and ML drift after vendor updates; separate bundled surveillance modules for distinct review.","delivery":"Delivery framing: Delivery centers on legal analysis covering Equal Protection, Title VI, and Loomis-line doctrine; Community Advisory Committee stand-up; a parallel-run pilot lasting six to twelve months; and override-pattern supervision by demographic group. Acceptance criteria include producible case-specific rationales on contest, disparity metrics within agreed bands, and documented appeals. Risks include cementing biased baselines via automation bias and vendor lock-in that blocks later audit."},"roleDetails":{"sales":{"customerProblem":"State DOCs need more consistent, auditable custody/housing decisions without locking in disparate impacts or unauditable proprietary scores.","stakeholders":["DOC director / classification leadership","Facility wardens","General counsel / civil rights counsel","Formerly incarcerated advisors","IT/OMS owners"],"discoveryQuestions":["Can the vendor provide disaggregated accuracy by race, gender, and age as a contract term?","On individual contests, can the agency produce factors and weights for that person's score?","What data-portability and audit rights survive termination?","How will overrides be justified both when accepting and rejecting recommendations?"],"valueHypothesis":"Bounded classification-governance design can improve consistency and contestability relative to paper/point systems if Title VI testing and human review are enforceable—not if the tool is a black box.","potentialEngagement":"Contract-redline and pilot-design engagement producing a Careful Implementation checklist and go/no-go metrics before RFP award.","unsupportedClaims":["Guaranteed reduction in violence or recidivism","That Loomis-style warnings alone satisfy due process aims","Fairness without routine disaggregated outcome tracking","Equivalence across actuarial vs opaque ML products"]},"engineering":{"fit":"Fit for DOC environments that can parallel-run classification and instrument demographic overrides; poor fit if auto-accept or surveillance modules are bundled without separate assessment.","architectureIntegration":"Classification service consuming OMS/EHR/program data; rationale API for contests; override journal; drift monitors on score distributions.","prerequisites":["Clean data mapping across source systems","Title VI disparate-impact testing commitment","Independent validation rights","Documented human-review and override workflow"],"constraints":["Proprietary algorithms blocking meaningful contest","Unobservable counterfactual ground truth for violence","ML components may drift after vendor retrains","State algorithmic-decisioning laws may add duties"],"security":"Protect sensitive correctional and health data; CJIS; prohibit silent repurposing for discipline/surveillance without new DPIA-equivalent review.","proofOfValue":["Calibration and PPV/NPV by demographic group","Misconduct rates by classification level vs baseline","Appeal outcomes and override-justification quality audits","Parallel-run comparison to current process"]},"delivery":{"work":"Legal analysis, CAC formation, pilot parallel-run, supervisor training on demographic override review, public reporting.","dependencies":["Counsel capacity for Title VI/equal-protection analysis","Compensated community advisors","OMS data quality","Vendor cooperation on rationales"],"ownership":"Named DOC official for classification outcomes and disparate-impact monitoring.","skillsAdoption":"Placement officers trained on score interpretation and automation-bias resistance; supervisors on demographic override patterns.","governanceCheckpoints":["Pre-contract mitigation on contest/disparate impact","Pilot mid-point disparity review","Annual substantial-risk reassessment","Triggered reassessment on model migration"],"acceptanceCriteria":["Case-specific rationales producible on sample contests","Disparity metrics within pre-set bands or deployment frozen","No auto-accept configuration in production"],"risks":["Automation bias cementing poor baselines","Vendor lock-in blocking audits","Title VI complaints if monitoring is weak","Expansion into predictive-violence modules without new review"]}},"retrievedAt":"2026-09-17T03:15:10Z","enrichedAt":"2026-09-17T03:20:00Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Requires sustained training for placement officers and compensated community participation; accessibility of appeals and understandable rationales is a rights requirement, not optional UX.","procurementImplications":"Negotiate Level 1/2 terms: disaggregated accuracy disclosure, independent audit at renewal, data ownership/portability, Title VI testing, liability/indemnification push, termination with data exit.","operatingModelImplications":"Classification becomes a monitored socio-technical process with CAC oversight and annual reassessment, not a back-office scoring upgrade.","sourceVerification":{"openedUrl":"https://counciloncj.org/assessing-ai-in-action-a-case-study-on-the-use-of-ai-powered-classification-tools-in-state-corrections/","referenceExcerpt":"Classification decisions have direct liberty impacts across custody level, housing, and programming eligibility.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}