Editorial method

Evidence before intensity.

Cognitive threats are easy to exaggerate because technical capability, public fear, and actual behavioral effect are often collapsed into one claim. This project keeps them separate.

01

Classify the claim

Each material statement should be identifiable as verified fact, documented allegation, participant claim, expert interpretation, inference, dispute, unknown, or scenario.

02

Grade the source

Prefer primary records, court documents, statutes, official audits, peer-reviewed research, direct technical evidence, and transparent datasets. Secondary synthesis is used with attribution.

03

Separate capability from effect

A system that can generate persuasive language has not thereby proven durable behavior change. Reach, authentic exposure, causation, and long-term effects require separate evidence.

04

Preserve counterevidence

Negative findings, failed replications, low engagement, false positives, and benign uses remain visible. They are not treated as obstacles to the narrative.

05

Minimize operational detail

Research on grooming, influence operations, synthetic personas, or coordinated networks is framed for recognition, governance, and defense—not as a blueprint for abuse.

06

Reverify unstable facts

Current laws, office-holders, product features, platform policies, model capabilities, sanctions, cases, and active conflicts must be checked again before a formal publication release.

07

Correct publicly

Substantive corrections should state what changed, why it changed, and what evidence supports the revision. Silent rewrites are avoided for consequential claims.

PUBLIC LABELS

Evidence states used throughout the site

Documented

Supported by public records, audits, court filings, peer-reviewed studies, or confirmed deployments.

Emerging

Demonstrated in controlled research, limited deployments, or credible early evidence; scale and durability remain uncertain.

Contested

Evidence is mixed, causal claims are disputed, or commercial claims exceed independent validation.

Scenario

A defensive projection used to test resilience. It is not presented as a current event or proven capability.

CURRENT LIMITATION

Research is supplied, not independently peer reviewed by the site.

The first release preserves the supplied reports and their bibliographies. It adds editorial labels and cautions, but it does not claim to have independently validated every footnote or current legal statement.

CORRECTION PATH

Evidence can be challenged.

Submit a source correction, methodological objection, missing counterexample, or legal update through the public participation form. Include a stable source and the exact claim affected.

Submit a correction

Search the archive

ProvenanceDependencySurveillance