WriteVoice
Dial · Do you pin things down, or speak in general truths?

Broad to Grounded

Broad to Grounded is the WriteVoice dial for specificity: whether your claims are tied to a when, a where and a who, with real numbers, named sources and events in order, or framed as general truths and advice that would read the same anywhere.

Published

What each end sounds like

Example sentences written by WriteVoice for illustration.

Broad end · The PhilosopherTravel is best when you leave room for the unexpected. Plans are useful, but the moments people remember are usually the ones nobody scheduled.
Grounded end · The Field ReporterIn May we got stuck in Porto for two extra days because the 14:10 train was cancelled. Best 48 hours of the trip: a €9 lunch at a place called Gazela and a free fado night.

What this dial looks for

  • Whether your claims come with a when, a where and a who: dates, places, people, real numbers and named sources.
  • Whether events are told in the order they happened, as things you could check.
  • At the other end: general truths and advice framed to hold anywhere, any time.

Your score is a percentile among people, not a grade: 50 is the middle of the human range. Under 150 words it is shown as a rough range. Your raw writing is never shown on public pages.

What the research found

  • Specificity is the dial that held up best in our research, and the one where readers most easily agree with the score when given its description.
  • Most chat models lean general, especially compared with people posting in online discussions; some recent Claude models are as specific as people, or more.

From WriteVoice research comparing people and AI models answering the same prompts. Default AI sits at 43/100 on this dial. More on the research.

What AI does with this dial

Most chat models lean Broad (general advice in neat points); some Claude models lean Grounded. Your default AI's position is marked on your result.

See also the tells of AI writing and how to make AI write like you.

Voice types on this dial

References

  • WriteVoice research (2026): the specificity dimension of human vs AI writing.
  • Biber, D. (1988). Variation across Speech and Writing. Cambridge University Press.

Frequently asked questions

What does the Broad to Grounded dial measure?

Broad to Grounded is the WriteVoice dial for specificity: whether your claims are tied to a when, a where and a who, with real numbers, named sources and events in order, or framed as general truths and advice that would read the same anywhere.

How does WriteVoice read Broad to Grounded?

From your own writing. An AI model reads your answers against a fixed rubric from our research and your result is placed among people answering the same questions, so 50 is the middle of the human range. It looks at: whether your claims come with a when, a where and a who: dates, places, people, real numbers and named sources; Whether events are told in the order they happened, as things you could check; At the other end: general truths and advice framed to hold anywhere, any time.

Where does AI sit on Broad to Grounded?

Most chat models lean Broad; some recent Claude models lean Grounded. Specificity is the dial that held up best in our research, and the one where readers most easily agree with the score when given its description. Most chat models lean general, especially compared with people posting in online discussions; some recent Claude models are as specific as people, or more.

Is it better to be broad or grounded?

Neither. Broad to Grounded describes a habit, not a quality. Each end suits different readers and situations; the useful thing is knowing where you sit, so you can choose when to move.

Where do you sit on Broad to Grounded?

Chat for three minutes or paste something you wrote. Get your voice type, your five dials and where default AI sits on each. Free.

The other 4 dials