Category 11 of 12 · Environment level
Institutions use data, machine learning, simulation, or forecasting to predict collective behavior and guide interventions affecting groups or populations. Evidence trail for AIP-11-0418: 10 selected sources, 3 with a bounded automated source check.
Definition and boundary
Forecasting of protests, conflict, migration, crime, public health, markets, voting, compliance, or other collective behavior to guide intervention. Evidence trail for AIP-11-0419: 10 selected sources, 3 with a bounded automated source check.
Ordinary descriptive statistics that do not predict or direct intervention, and aggregate planning systems that cannot be used to identify individuals may present a substantially different risk. Evidence trail for AIP-11-0420: 10 selected sources, 3 with a bounded automated source check.
Strategic and human significance
Forecasts can become self-fulfilling when intervention creates the data that validates the model. Evidence trail for AIP-11-0422: 10 selected sources, 3 with a bounded automated source check.
High-risk labels can shift the presumption from individualized evidence toward statistical suspicion. Evidence trail for AIP-11-0423: 10 selected sources, 3 with a bounded automated source check.
The same architecture can support aid planning or preemptive repression depending on the unit of analysis, data, intervention, and due-process safeguards. Evidence trail for AIP-11-0421: 10 selected sources, 3 with a bounded automated source check.
Change from pre-AI practice
Evidence maturity
Aggregate forecasting supports humanitarian and conflict planning, while individual risk scoring and rare-event prediction often produce severe error, bias, and feedback loops.
Systems such as ViEWS and Project Jetson forecast conflict or displacement for planning and humanitarian response. Evidence trail for AIP-11-0429: 10 selected sources, 3 with a bounded automated source check.
Predictive policing and surveillance systems have scored people and places for intervention. Evidence trail for AIP-11-0430: 10 selected sources, 3 with a bounded automated source check.
Existing conflict continuation can be more predictable than rare novel events; local individual prediction often has poor validity. Evidence trail for AIP-11-0431: 10 selected sources, 3 with a bounded automated source check.
Conflicting findings and scope boundaries
These are not errors to hide. They reflect different study designs, measures, time horizons, and operational contexts.
TENSION-LAW-01
Evidence position A: The reports identify existing enforcement actions and regulatory controls relevant to synthetic media, sensitive data, manipulation, and automated decision systems.
Evidence position B: Case posture, effective dates, definitions, constitutional constraints, and remedies vary across jurisdictions and change over time.
Publication rule: Date-stamp legal statements, identify the jurisdiction and procedural posture, and require current primary-source recheck before publication acceptance.
TENSION-LEGAL-CURRENTNESS-01
Evidence position A: The reports identify statutes, regulations, enforcement actions, and litigation relevant to manipulative or synthetic AI systems.
Evidence position B: Effective dates, amendments, injunctions, appeals, definitions, and remedies vary by jurisdiction and can change after the report date.
Publication rule: Require source-specific currentness review by an authorized human before publication acceptance.
Conceptual mechanisms
Event and trend forecasting. Evidence trail for AIP-11-0432: 10 selected sources, 3 with a bounded automated source check.
Network and mobility analysis. Evidence trail for AIP-11-0433: 10 selected sources, 3 with a bounded automated source check.
Risk scoring of people or places. Evidence trail for AIP-11-0434: 10 selected sources, 3 with a bounded automated source check.
Agent-based or scenario simulation. Evidence trail for AIP-11-0435: 10 selected sources, 3 with a bounded automated source check.
Intervention feedback into future training data. Evidence trail for AIP-11-0436: 10 selected sources, 3 with a bounded automated source check.
These descriptions explain capability and risk. They intentionally omit procedures, targeting criteria, scripts, and evasion methods.
Evidence and examples
Evidence trail for AIP-11-0437: 3 selected sources, 1 with a bounded automated source check.
Sources: report ref. 20: ViEWS: A political violence early-warning system - ResearchGate (opens in a new tab), report ref. 24: ViEWS: A political Violence Early Warning System (opens in a new tab), report ref. 25: ViEWS: A political violence early-warning system | Journal of Peace Research (opens in a new tab)
Evidence trail for AIP-11-0438: 2 selected sources, 1 with a bounded automated source check.
Sources: report ref. 21: CITY OF CHICAGO OFFICE OF INSPECTOR GENERAL ADVISORY CONCERNING THE CHICAGO POLICE DEPARTMENT'S PREDICTIVE RISK MODELS (opens in a new tab), report ref. 22: Evaluation of the Chicago Police Department's Strategic Decision Support Centers - RAND (opens in a new tab)
Case studies show documented events or bounded experiments. They do not establish prevalence, general causation, or guaranteed persuasive effect.
Risk and failure analysis
Detection and defense
Caveat: Auditing requires comparing enforcement data with independent measures of underlying events. Evidence trail for AIP-11-0447: 10 selected sources, 3 with a bounded automated source check.
Caveat: A simple interface is not itself proof of an invalid model, but it increases misuse risk. Evidence trail for AIP-11-0448: 10 selected sources, 3 with a bounded automated source check.
Caveat: Such proxies require strong legal scrutiny and contextual review. Evidence trail for AIP-11-0449: 10 selected sources, 3 with a bounded automated source check.
Governance and safeguards
Open questions
Sources and evidence boundary
Primary synthesis: Predictive Population Management Research. The complete report is retained in a non-public provenance directory with SHA-256 07aec0936a99a11bd93706fad9af992235c4c40608a5db25962752648c16560d.
Thirty high-impact references received bounded automated retrieval, official corroboration, or stronger-source substitution. All 91 selected references are used by the 501-claim citation graph, but the full report corpus and human editorial acceptance remain unverified. Source type labels are editorial classifications, not quality scores.
Metadata is resolved, but the linked source has not received this release's independent content-scope check.
Human Rights Watch describes its reverse engineering and analysis of the Integrated Joint Operations Platform application and its targeting implications.
Limits: The source is a human-rights investigation rather than an independent technical audit commissioned by the system operator.
Metadata is resolved, but the linked source has not received this release's independent content-scope check.
Metadata is resolved, but the linked source has not received this release's independent content-scope check.
Metadata is resolved, but the linked source has not received this release's independent content-scope check.
The Chicago Inspector General advisory supports bounded claims about governance, validation, transparency, and operational concerns surrounding predictive risk models.
Limits: The findings concern specific Chicago Police Department programs and should not be generalized to every predictive system.
Metadata is resolved, but the linked source has not received this release's independent content-scope check.
Metadata is resolved, but the linked source has not received this release's independent content-scope check.
The ViEWS article supports a bounded claim that transparent statistical systems can forecast forms of political violence at aggregate geographic and temporal scales.
Limits: Performance varies by event type, horizon, geography, and base rate. Aggregate early warning does not justify identifying or coercively targeting individuals.
Metadata is resolved, but the linked source has not received this release's independent content-scope check.