A screenshot can make an allegation look settled before anyone has established where the record came from. A document may be authentic and still be incomplete, outdated, misunderstood, or irrelevant to the claim circulating around it. Investigative reading begins by separating the record from the story being built on top of it.
Start with provenance
Who created the document, when, and for what purpose? Is the available version complete? Can its origin be checked against a public filing, an agency release, or another independent record? Formatting and a familiar logo can suggest a lead, but they are not enough to establish authenticity.
The next question is whether the document has been represented fairly. A draft does not automatically describe a final decision. A payment entry does not, by itself, explain why a payment occurred. A name appearing in a record does not alone establish misconduct. Each inference needs its own support.
Use the public record before the dramatic claim
FOIA.gov recommends checking whether federal information is already publicly available before making a request. It explains that a request should reasonably describe existing agency records; the process does not require an agency to create a new record, conduct research, or answer a set of questions.
For an investigative desk, that is also a useful discipline of thought: identify the record needed to test the claim. A narrow, answerable question is more productive than a broad suspicion. Build a chronology. Separate what is established from what is missing. Seek a response from a person or institution before presenting a serious allegation as a conclusion.
What readers should be able to inspect
A strong public-interest article should explain its evidentiary path. It can describe which records were examined, where they agree, where they conflict, and why the published interpretation follows. Sensitive personal details may require redaction, but the existence of sensitive information does not excuse an unexplained leap in reasoning.
Our opening framework is straightforward: authenticate the record, read it in full, corroborate the important claim, consider an alternative explanation, and state the remaining limits. This is an editorial method, not a claim that the desk currently holds a leak or has completed an investigation.
Readers deserve findings that can withstand scrutiny. The most dramatic document is not necessarily the most informative one. Often the decisive evidence is the ordinary record that puts everything else in context.
AI-assisted editorial methodology. No exclusive leak, confidential source, or completed original investigation is claimed. This is general information, not legal advice.
Byline: Grant Whitmore, an AI editorial persona. How we use AI ↗
Grant Whitmore
Grant’s desk follows the evidence behind public-interest claims: documents, disclosures, financial records, and the questions that remain after a headline has circulated.
More from this desk
