Cases on Record

Editorial policy

Last updated

Cases on Record explains U.S. public records in plain English. Every article starts from a public record and links to it.

What we write about

We cover three areas: auto insurance and accidents, injury lawsuits, and debt and loans. Articles are based on primary records — court opinions, the Consumer Financial Protection Bureau’s complaint database, and the National Highway Traffic Safety Administration’s recall and owner-complaint records — and, where helpful, on news reports that point to those records.

How an article is made

  1. A record is selected. Each day our system groups new records into stories and picks one that has at least two sources, including at least one primary record.
  2. Facts are listed with their sources. Every fact we may use is written down with the link it comes from. Counts and dates are calculated directly from the records, not estimated.
  3. A draft is written with AI assistance. A language model running on our own computer writes a first draft using only those listed facts. It is instructed not to add facts, round numbers, speculate, or give advice.
  4. Automated checks run. Before a person sees it, the draft is checked so that every number and name traces to a listed source, no sentence copies the source wording, no private individual is named, no advice or guarantees appear, and it does not repeat an earlier article.
  5. An editor reviews and approves. A person reads every draft and either approves it, edits it, or rejects it. Nothing is published without that approval.

What we do not do

Sources and attribution

Each article lists its sources with links and the date they were checked. See our sources page for the record systems we use and how we credit them.

Corrections

If we get something wrong, we correct it and note the change. See our corrections policy.