As AI systems become better at producing fluent and convincing legal, administrative and civic analysis, a basic question becomes more important: can we still tell what is actually supported by evidence, what comes from a source, what is an inference, and what remains only an assumption?
The new Civic Forensics White Paper 2026 — “When Persuasive Text Is Not Evidence” addresses this problem through a methodological framework built around Evidence Ceiling, Claim Drift and Claim Provenance, supported by the control chain:
CLAIM → SOURCE → STATUS → CONFIDENCE → ACTION
The paper argues that AI literacy should include evidence literacy: the ability to distinguish sources, claims, inferences, uncertainty and missing evidence when working with AI-assisted analysis.
The central principle is simple: better AI does not reduce the need for evidence discipline — it makes it more important.
Read the White Paper and explore the Civic Forensics methodology: