Claim detection

Large language models are trained to recognize linguistic patterns, and to use those patterns to produce new content. However, they aren't trained to recognize whether the content they’re being trained on is true. When Writer suggests a quote or a stat, the model is suggesting that this is a statement that'd fit well in the content, but it has no way to confirm that the statement is factual. A human editor should verify any facts or statistics you see before publication. 

What is claim detection mode?

With claim detection mode feature, content that might need to be fact-checked will be automatically highlighted so you can more easily identify, verify, and edit (if necessary) any potentially problematic claims.

How do I access and use claim detection mode?

Go to app.writer.com and open up a doc. Once in the document, select the claim detection mode button (which looks like a shield) to highlight any claims that Writer can't verify.

FAQs

Q: What does Writer consider to be a "claim"?

A claim can be a statistic, a quote, or an anecdote.