Being open about AI in your writing may hinge less on the tool than on why people skip saying so.

This explores the practical side of AI disclosure: what tools and habits help writers say when and how AI shaped their text. The collection has little on disclosure tools as such. What it does have is one promising design direction, one detection-style approach, and a clear picture of why people don't disclose. That last part may matter more than any tool.

The most direct tool evidence comes from shared writing editors. When pairs of writers could choose how much of each other's AI prompting to see, they strongly preferred more visibility. Seeing when, how and where a collaborator used AI helped them understand the collaborator's thinking and check AI-generated passages. Some found full sharing intrusive and felt self-conscious, though, which suggests disclosure works better as an adjustable dial than as an on/off switch Do writers want to see each other's AI prompts in shared editors?. A second approach records the writing process instead of relying on a statement. AI contributions tend to arrive in sudden bursts that break a writer's normal rhythm, so process logs can reliably flag wholesale handoff to AI. Ordinary back-and-forth help, however, looks almost the same as writing with little AI Can process data distinguish AI delegation from ordinary collaboration?. Process data can catch the extreme case, but it can't stand in for honest labeling of everything in between.

The bigger obstacle is psychological. People say they don't feel they own AI-generated text, yet they still don't credit the AI publicly. They treat it like an invisible ghostwriter, even when the text was personalized for them Do people feel they own AI-generated text they use?. Readers and writers also disagree about what needs disclosing. Readers judge disclosure more necessary than writers do, especially when AI text was pasted in directly and couldn't easily be replaced, and how much effort the writer put in made no difference Do readers and writers differ on AI disclosure necessity?. A useful rule of thumb follows: disclose based on how much of the AI's text survives into the final piece, not on how hard you worked. That matters because writers edit AI paragraphs only about 23% of the time, and lightly when they do Do writers actually edit AI-generated text before publishing?.

The cost of disclosing is smaller than writers seem to fear. Adding an AI disclosure to a news article lowered ratings from both human and LLM raters by less than 0.15 points on a 7-point scale Does disclosing AI assistance make readers trust articles less?. Readers with higher AI literacy showed smaller drops still, and some reacted positively Does AI literacy reduce the damage from AI disclosure?. Not disclosing has its own costs. Unfounded accusations of AI use already fall on human writers, and those accusations work more as gatekeeping than as accurate detection Do unfounded AI accusations harm human writers instead?. A clear, voluntary statement gives writers some control over that suspicion.

The surprising connection: professional writers locate authenticity in the process of making a piece, not only in the finished text Where do writers locate authenticity in AI co-writing?. If that's where authenticity lives, the most fitting disclosure practice may be to describe the process rather than attach a yes/no label. For example, a writer could say the AI helped with brainstorming, and that the drafting was their own. The creative stages that show up in co-writing studies, ideation, organizing and drafting, give writers a ready vocabulary for that kind of statement How do writers use AI through different creative stages?.