GLOSSARY · GHOSTWRITING
AI disclosure
AI disclosure is the practice of saying, where it matters, that AI tools were used to draft, edit or generate a piece of content. Norms are still forming. Few readers expect a note on a LinkedIn post that was drafted with AI and then substantially rewritten by the author, just as nobody discloses a spell checker. Readers do react badly to content that presents machine-generated text as first-hand experience, and some publications, platforms and clients now have explicit rules. The practical rule is that the named author must stand behind every claim as true and every opinion as their own, whatever tool produced the words.
How it works in practice
Three questions decide what disclosure a piece of content needs. First, who is responsible? If a founder uses an AI tool to draft and then edits until every sentence is something they would say, the founder is the author and the tool is an instrument. If a post is generated and published unread, nobody is the author, and the lack of disclosure is the least of the problems. Second, what does the reader assume? A personal post implies personal experience. If the "story about a client last week" was produced by a model and never happened, that is deception regardless of any AI label. Third, what do the rules say? Some newsletters, publications, employers and clients require a disclosure line; LinkedIn itself labels some AI-generated images and has policies on misleading content.
Follow the rules of the place you publish. Beyond those, the honest minimum is this: never claim experience you did not have, never publish a figure you did not check, and if asked whether AI was involved, answer truthfully. Many writers find a short note on their profile or in a newsletter's about page is enough: "Drafted with AI tools and edited by me; every claim is mine." The bigger reputational risk is not using AI; it is sounding like it, which readers detect from stock openers, generic lists and stories without specifics.
Example
A founder's newsletter about page says: "I draft with AI and rewrite everything in my own words. Every number has a source I can show you, and every story is one I was part of." Individual posts carry no label.
Common mistake
Publishing AI output with first-person anecdotes left in. The founder is asked about a conversation that never happened, cannot answer, and the audience now has to wonder about every other post.