AI Use Policy.

AI Use Policy

Last updated: May 26, 2026

VTScout uses artificial intelligence (AI) tools in some parts of our publishing workflow. We believe readers deserve clear, specific information about where AI is involved, what it is used for, and what safeguards we apply. This page makes our practices explicit. See also our Editorial Standards and Ethics Policy.

Where AI is used

• Drafting from public records. Press releases from law-enforcement agencies, court filings, and other public-record sources may be summarized and rewritten with AI assistance to produce an initial draft. Every such draft is reviewed by a human editor before publication.
• Headlines and SEO metadata. AI may suggest headlines, meta descriptions, and tag candidates. Editors choose the final wording.
• Image generation. When we cannot obtain or appropriately use a documentary photograph for a story, we may use AI-generated illustrations. These images are always visibly labeled and machine-readable (a role=note disclosure) so readers know the image is illustrative, not a photograph of the actual scene or people.
• Alt text for accessibility. Images uploaded to the site may have descriptive alt text generated by an AI service (OpenAI Vision). Editors can override the generated text at any time. See our Subprocessors page for vendor details.
• Proofreading. AI tools may flag typos, grammar issues, and clarity problems. Editorial judgment makes the final call.

Where AI is not used

• Originals. Stories tagged “Original reporting” are written, edited, and verified by humans without AI rewriting.
• Quotations. AI is never used to generate quotes attributed to real people. Quotes come from named sources, transcripts, or named press releases.
• Verification. Facts are verified against primary sources by humans. AI is not the last word on whether a claim is accurate.
• Sensitive coverage. Stories involving minors, sexual-assault victims, deaths, or fast-developing public-safety events receive additional human review and do not auto-publish.

Disclosure on every article

Every AI-assisted article carries an “About this report” disclosure block at the bottom that names the source material and indicates AI assistance. Original reporting carries a distinct “Original reporting” badge instead.

Source guardrails

Our publishing pipeline includes a “Source Guard” gateway that blocks ingestion of certain copyrighted publishers, routes others to human review, and applies a police-framing validator to catch the specific failure mode of mislabeling a non-police story as a police story. These guardrails sit in front of every automated publish and are enforced at the REST insertion layer.

What we do not do

• We do not republish AI-rewritten copies of another outlet’s copyrighted reporting.
• We do not synthesize photographs of named individuals.
• We do not use AI to fabricate sources, statistics, or events.
• We do not present AI-generated content as the work of a named reporter.

If AI gets something wrong

When an AI-assisted story contains a factual error, the correction process is identical to any other correction — see the Corrections page. We log the error class so we can adjust our workflows.

Reader questions

For questions about how AI is used at VTScout, email [email protected] with “AI Policy” in the subject line.