One-sentence brief
Data described as “only metadata” can become highly consequential when combined across sources and used in eligibility, policing, employment, housing, insurance, or advertising decisions.
INFORMATION INTEGRITY, OVERSIGHT, AND DIGITAL RIGHTS
A rights-centered data-flow model showing how timestamps, locations, device identifiers, purchases, public records, and other signals can be aggregated, brokered, and used to infer identity or sensitive attributes.
ORIENTATION
Data described as “only metadata” can become highly consequential when combined across sources and used in eligibility, policing, employment, housing, insurance, or advertising decisions.
WORKING BRIEF
Record what was observed, who collected it, the user-facing relationship, purpose, consent or legal basis, retention period, and whether collection was direct, inferred, or acquired.
Location, time, device, purchase, public-record, and communication signals can be linked across brokers and platforms. The combined profile can reveal more than any single record.
An inferred identity, health interest, religion, political leaning, economic status, or risk score is not the same as a self-declared fact. Record model, data, confidence, error, and contestability.
Repeated licensing and resale make it difficult to trace original collection, purpose, correction, deletion, and downstream recipients. Public diagrams should show those handoffs.
The same signal can be used for advertising, fraud prevention, housing, employment, law enforcement, or eligibility. Explain the decision consequence and human review path.
Access, correction, deletion, opt-out, explanation, appeal, audit, regulator complaint, civil action, and policy change are distinct forms of contestability.
COMPLETE DOSSIER
Terms are defined for this site’s evidence method, not as universal legal or clinical definitions.
| Layer | Evidence that supports it | What it does not establish | Safer public treatment |
|---|---|---|---|
| Observation | Archived item, timestamp, text, image, filing, or public record | Identity, sponsor, intent, reach, or effect | State exactly what was observed |
| Attribution | Direct technical, documentary, legal, or independently corroborated links | Organization-wide responsibility or hidden principal without evidence | Name the claimant and confidence |
| Impact | Measured exposure, outage, decision, or adjudicated harm | Causal effect from virality or visibility alone | Separate reach from effect |
| Correction | Versioned status, reason, affected routes, and preserved prior state | Silent certainty upgrade or deleted uncertainty | Publish the change and reopening trigger |
Are observation, behavior, actor, sponsor, intent, reach, effect, legal finding, and remedy separated?
Pass condition: Each asserted rung has its own evidence and status.
Do multiple citations trace to genuinely independent origins?
Pass condition: The citation tree identifies common origins, actor-controlled material, and circular reporting.
Does the wording match the actual evidence and decision forum?
Pass condition: Claimant labels, confidence, disputes, unknowns, and safer verbs remain visible.
Can a later record revise one claim without silently rewriting the rest?
Pass condition: Version, reason, affected routes, superseded state, and reopening trigger are recorded.
Does the page avoid operational enablement, identity harm, profiling, and protected-trait inference?
Pass condition: Only public-interest educational analysis is present, with privacy and remedy boundaries.
RESEARCH EDITION
This page follows the public method for provenance, confidence, source independence, alternative accounts, limitations, review state, and visible correction.
CONTINUE