Proof standard
The source, claim, data transformation, artifact, editorial acceptance, accessible delivery, correction path, and reader evidence remain attributable.
Newsrooms and public explainers
For data journalists, graphics editors, newsroom developers, and public-interest communicators evaluating AI-assisted analysis and visual explanation.
Decision to support
Proof standard
The source, claim, data transformation, artifact, editorial acceptance, accessible delivery, correction path, and reader evidence remain attributable.
Working output
A publication and correction plan that names what AI may assist, what editorial authority must inspect, and what readers still need tested.
Composed research
These are the original sections from the visual research experiences, not summaries or reference cards. Expand each step to read it here, or open its stable page for direct linking.
Begin with the public claim, intended reader, stakes, and delivery setting.
Open this section as its own page →Distinguish comprehension, confidence, accessibility, trust, and later use.
Open this section as its own page →Design review around demonstrated semantic, visual, interaction, and critic failures.
Open this section as its own page →Keep editorial acceptance separate from mechanical checks and model judgments.
Open this section as its own page →Preserve publication, correction, recheck, and ownership as separate events.
Open this section as its own page →Ask before acting
Can a reader or editor trace every consequential claim back to its source and transformation?
Has the exact delivered artifact been checked across devices, access paths, and interaction states?
Who owns correction, republication, disclosure, and reader-facing notice after a material change?
Focused analysis
Named evidence