Writers admit AI help most freely to close collaborators and least to strangers, and readers mind hidden AI most in personal writing.

This explores whether the rules for admitting AI helped you write shift with the audience, from a co-author or a friend to anonymous public readers. No study in the collection compares private and public writing head to head. Read together, though, these notes show a clear pattern: people are most open about AI use in small private circles and least open in public, while readers are most bothered by hidden AI use in the most personal kinds of writing.

In private collaboration, openness is the norm. When pairs of writers shared an editor, they consistently preferred to see more of each other's AI prompting. Knowing when and how a partner used AI helped them follow the partner's thinking and check the generated text. Some found full visibility a bit self-conscious, but they still preferred it Do writers want to see each other's AI prompts in shared editors?. In public the pattern flips. People say they don't feel like the owners or authors of AI-generated text, yet they don't credit the AI when they publish it. They treat it as an invisible ghostwriter Do people feel they own AI-generated text they use?. Having more control over the output makes people feel more ownership Does user control over AI text shape feelings of ownership?, which may help them feel entitled to leave the AI uncredited.

Writers may have a reason to stay quiet, and it is strongest where you might least expect. When readers learn AI was involved, they rate the writer as less trustworthy, less caring, and less likable. The biggest drops come in interpersonal writing, because readers see AI as unable to genuinely empathize, so using it for a personal message feels like breaking a social expectation How does revealing AI authorship change reader trust?. So the private-versus-public line isn't the only one that matters. A personal note sent to one person may carry a higher disclosure cost than a public report. Readers also think disclosure is more necessary than writers do, especially when AI text goes into the final piece directly and couldn't easily be replaced. How much effort the writer put in doesn't change that judgment Do readers and writers differ on AI disclosure necessity?. Readers with more AI literacy penalize disclosure less, so the cost depends partly on who is reading Does AI literacy reduce the damage from AI disclosure?.

There's also a structural reason public settings are different. AI writes for the person typing the prompt, not for a broader audience it has imagined. When that text is published, it reaches readers the model never had in mind. Text written like a private reply ends up presented as public authored writing Does AI writing collapse the author-to-public relationship?. This matters because writers rarely change AI drafts. In one study they edited only 23% of AI paragraphs, and the edited versions stayed about 96% similar to the original Do writers actually edit AI-generated text before publishing?. Readers then pick up shifts in tone: AI help made writers seem more confident, more extreme, and more privileged across all 29 traits measured Does AI writing assistance change how readers perceive the writer?. Undisclosed AI in public writing therefore changes how the author comes across, not just who gets credit.

When disclosure fails in public, people don't simply stop caring. They start suspecting. Accusations of AI use in public comment threads often target text that shows no real signs of AI, so the accusation works more like gatekeeping than detection, and human writers can be wrongly discounted Do unfounded AI accusations harm human writers instead?. Hiding AI use publicly can therefore end up hurting honest writers who used none.