ChatGPT is the tool most people reach for first, so it sits behind much of the AI text that teachers, editors and hiring managers see. This page covers what its drafts tend to look like, what changes once someone edits them, and how to read a Winston AI result without jumping to conclusions.
Why ChatGPT text turns up everywhere
Students use it to outline essays, job seekers use it to polish cover letters, and office teams use it to answer email faster. Because it is the default for so many people, a suspicious paragraph in a classroom or inbox is more likely to have started in ChatGPT than anywhere else. That doesn't change how you check it. Winston AI looks for machine-written text in general and reports how likely it is, not which tool produced it.
What happens when people edit a draft
Few people paste a raw answer anymore. More often they ask for a draft, swap in their own examples, cut the summary paragraph and fix a few phrases. That mix shows up in the results. Untouched passages tend to stay red, rewritten sentences often turn yellow, and lines the writer added from scratch usually carry no highlight at all.
- Light edits (a few swapped words) usually leave the overall score high.
- Heavy rewrites spread the signal out, so look at which sentences stay red rather than at the headline number.
- Paraphrasing tools tend to produce yellow, mixed-signal sentences. Our guide on whether paraphrasing hides AI writing goes further.
Before you act on a score
A high score is a reason to look closer, not a verdict. Ask for an outline or an earlier draft, check version history, compare the piece with the person's past work, or simply talk it through. Students who wrote their own work can usually explain how they got there. If you teach, the walkthrough on spotting a ChatGPT essay pairs well with a scan.