Do co-writers want to see each other's AI prompts, even if having their own on display makes them self-conscious?

This explores whether letting co-writers see your AI prompts, and not just the finished text, makes collaboration better. The most direct evidence is a study of sixteen writing pairs in shared editors: writers clearly preferred tools that showed more prompting activity, because it let them see when, how and where a partner had used AI Do writers want to see each other's AI prompts in shared editors?. They gave two reasons. Seeing a prompt helped them follow a partner's thinking, and it told them which passages were AI-generated and should be checked. The same study found a cost, though. Full visibility could feel intrusive, and some writers became self-conscious when their prompting was on display. So writers wanted to see their partners' prompts more than they enjoyed having their own watched. Keep in mind that this is a preference result. Nobody measured whether the shared drafts actually got better.

The checking benefit matters more than it first appears. Elsewhere in the corpus, writers edited AI-generated paragraphs only 23% of the time, and the edited versions stayed about 96% similar to the original Do writers actually edit AI-generated text before publishing?. That unedited text is not neutral. Across 29 measured traits, AI help shifted how readers saw the writer, making them seem more confident, more extreme and more privileged Does AI writing assistance change how readers perceive the writer?. In a shared document, visible prompts act as a flag for co-authors: this paragraph came from a model, so look at it twice. Without that flag, AI text can pass through two sets of eyes with neither person editing it.

There is also a less obvious reason prompts are worth showing. A prompt captures who the text was written for. AI writes for the person typing the prompt, not for the eventual audience a human author would have in mind Does AI writing collapse the author-to-public relationship?. In a collaboration, a partner who can read the prompt can see what the AI was asked to do and judge whether the result fits the shared goal. If they see only the output, they are reading text aimed at someone else's private request.

AI text can also be detected without anyone choosing to share. Process data shows that wholesale delegation to AI leaves a recognizable mark: text arrives in bursts that break from a writer's normal rhythm. Lighter collaborative help leaves no such mark Can process data distinguish AI delegation from ordinary collaboration?. Heavy AI use may come out anyway, so visibility matters most for the subtler help that no detector would catch. On whether disclosure seems necessary at all, writers and readers disagree. Readers rate AI disclosure as more necessary than writers do, especially when AI text goes in directly and can't be swapped out. How much effort the writer put in doesn't change their judgment Do readers and writers differ on AI disclosure necessity?. A co-author is partly a reader, which may explain why people want visibility from their partners.

What the corpus can't yet tell you is whether visible prompting changes how much ownership writers feel. Ownership of AI text rises with how much control the writer has over it Does user control over AI text shape feelings of ownership?. That suggests showing prompts could help each writer's contribution look like real steering rather than outsourcing. Nobody has tested that link, so treat it as an open question.