PREDICTIVE LAW ENFORCEMENT AND BEHAVIORAL THREAT ASSESSMENT
Fourteen source-bounded minimum-practice controls for individualized behavioral threat assessment and management.
ANALYTICAL & SAFETY BOUNDARIES
Analytical boundary: “Pre-crime” is an analytical label for an anticipatory decision architecture, not one technology or proof that future conduct can be known. Record retrieval is not prediction; identity matching is not prediction of conduct; association is not guilt; descriptive mapping is not necessarily forecasting; a concern category is not a calibrated probability; a score, tier, list, map, or alert is not lawful authority for coercive action; formal human presence is not meaningful review without time, evidence, independence, authority, and an effective intervention path; forecast accuracy is not demonstrated crime reduction; program silence is not proof of termination; a renamed successor is not proof of functional discontinuity.
Simulation safety boundary: Neutral, non-operational education. No real-person scoring, real targeting, weapons configuration, surveillance deployment, evasion, jamming instruction, or system-defeat guidance.
Source basis: Behavioral threat-assessment source reports and federal program evaluation material.
ORIENTATION
A checklist can structure inquiry, but it must not be confused with a validated probability model or used to turn diagnosis, disability, speech, ideology, grievance, or association into proof of threat.
WORKING BRIEF
The controls organize governance, evidence, support, review, correction, and exit. They do not calculate a probability of violence.
A human reviewer needs time, source access, independence, authority to reject, and an effective intervention path.
COMPLETE DOSSIER
No related page is required to understand this record.
Terms are defined for this site’s evidence method, not as universal legal or clinical definitions.
| Control | Minimum practice | Pass condition |
|---|---|---|
| Concrete concern, not population screening | A case begins with a documented safety concern rather than broad demographic, clinical, ideological, or association-based screening. | The initiating concern, source, time, and factual basis are recorded. |
| Individualized contextual inquiry | Assess the person, situation, environment, stressors, capabilities, protective factors, and alternative explanations in context. | The file distinguishes context from generalized profiles. |
| Defined multidisciplinary roles | Legal, investigative, safeguarding, clinical, educational, workplace, or community roles are defined without collapsing them. | Each participant’s authority and limits are recorded. |
| Separate observation, allegation, inference, and conclusion | Raw observations and source claims remain distinct from analyst interpretation and disposition. | The record exposes the transition between evidence states. |
| Source reliability and contradiction tracking | The team records provenance, credibility, corroboration, conflicts, missing data, and recency. | Contradictory evidence is retained rather than silently discarded. |
| Protective factors and alternatives | The review considers stabilizing relationships, services, changed circumstances, benign explanations, and reasons not to escalate. | Protective factors affect the management plan. |
| Proportional management, not label-only disposition | The response is tailored to the concrete concern and least coercive effective option. | The record explains why the chosen action is necessary and proportionate. |
| Services, safeguarding, and support | Service, voluntary support, conflict resolution, safety planning, and other non-enforcement options are available where appropriate. | Support is not represented as surveillance or proof of threat. |
| Independent supervisory review | A qualified reviewer can challenge the formulation, evidence, and proposed action. | Review is not a rubber stamp and includes documented disagreement. |
| Periodic reassessment | Risk and management are treated as dynamic rather than permanent labels. | Review dates and triggers are recorded. |
| Exit, correction, and deletion criteria | The system defines when a case closes, when records are corrected, what is retained, and what is deleted. | No indefinite retention by default. |
| Privacy-minimized information sharing | Only necessary, accurate, authorized information is shared with role-appropriate recipients. | Purpose, recipient, fields, and expiration are logged. |
| Accessible complaint and appeal | Affected people can raise errors or disproportionate consequences without inaccessible technical or procedural barriers. | The path, response time, and escalation route are public or otherwise accessible. |
| No certainty claim for rare violence | The team does not represent a concern category or checklist as a calibrated probability of attack. | Uncertainty, base-rate limits, and false-positive risk remain visible. |
A case begins with a documented safety concern rather than broad demographic, clinical, ideological, or association-based screening.
Pass condition: The initiating concern, source, time, and factual basis are recorded.
Assess the person, situation, environment, stressors, capabilities, protective factors, and alternative explanations in context.
Pass condition: The file distinguishes context from generalized profiles.
Legal, investigative, safeguarding, clinical, educational, workplace, or community roles are defined without collapsing them.
Pass condition: Each participant’s authority and limits are recorded.
Raw observations and source claims remain distinct from analyst interpretation and disposition.
Pass condition: The record exposes the transition between evidence states.
The team records provenance, credibility, corroboration, conflicts, missing data, and recency.
Pass condition: Contradictory evidence is retained rather than silently discarded.
The review considers stabilizing relationships, services, changed circumstances, benign explanations, and reasons not to escalate.
Pass condition: Protective factors affect the management plan.
The response is tailored to the concrete concern and least coercive effective option.
Pass condition: The record explains why the chosen action is necessary and proportionate.
Service, voluntary support, conflict resolution, safety planning, and other non-enforcement options are available where appropriate.
Pass condition: Support is not represented as surveillance or proof of threat.
A qualified reviewer can challenge the formulation, evidence, and proposed action.
Pass condition: Review is not a rubber stamp and includes documented disagreement.
Risk and management are treated as dynamic rather than permanent labels.
Pass condition: Review dates and triggers are recorded.
The system defines when a case closes, when records are corrected, what is retained, and what is deleted.
Pass condition: No indefinite retention by default.
Only necessary, accurate, authorized information is shared with role-appropriate recipients.
Pass condition: Purpose, recipient, fields, and expiration are logged.
Affected people can raise errors or disproportionate consequences without inaccessible technical or procedural barriers.
Pass condition: The path, response time, and escalation route are public or otherwise accessible.
The team does not represent a concern category or checklist as a calibrated probability of attack.
Pass condition: Uncertainty, base-rate limits, and false-positive risk remain visible.
RESEARCH EDITION
This page follows the public method for provenance, confidence, source independence, alternative accounts, limitations, review state, and visible correction.
CONTINUE