← All posts

Trends · August 19, 2026 · 3 min read

AI Disclosure Norms Are Forming Faster Than AI Policy

No law tells you whether to label an AI-assisted LinkedIn post. Audiences are settling the question anyway — here's the etiquette taking shape.

There is no rule that says you must disclose that AI helped write your LinkedIn post. There likely won't be one for years, and it's hard to imagine how such a rule would even be enforced. And yet, in the absence of policy, a norm is visibly forming — the way norms usually do, through thousands of small social verdicts about what feels honest and what feels like getting caught.

Etiquette moves faster than law

This is the normal order of things. Formal rules about communication technologies almost always trail the informal ones: communities settled expectations around ghostwriting, sponsored content, and photo retouching socially long before any regulator or platform wrote them down, and in most cases the eventual written rules just codified the etiquette that already existed. AI writing is running the same course, compressed. Regulatory attention so far has concentrated on the high-stakes zones — political ads, impersonation, synthetic media of real people — while the everyday question, do I tell people a model helped with this paragraph?, is being answered by audiences, one reaction at a time.

The distinction audiences are actually drawing

Watch how people respond when AI involvement surfaces, and a pattern emerges. The reaction rarely tracks whether AI was involved. It tracks what the AI was trusted with. The line that keeps re-forming, in comment sections and in private judgments, sits between assistance and substitution:

That's why "did you use AI?" is slowly becoming the wrong question, and "did you mean it?" is becoming the real one. A writer who feeds a model their own arguments and ships a draft they'd defend line by line is in a different moral category from one who typed a topic into a chatbot and pasted the output — even though both "used AI."

Why the calculus differs by platform

Norms are local. On a personal-brand platform like LinkedIn, the product being sold is you — your judgment, your experience — so undisclosed substitution reads as misrepresentation of the product itself. In contexts where writing was already understood to be collaborative or corporate — company blogs, newsletters with staff, executive comms — audiences never assumed a lone author, so AI involvement barely registers. The closer the content sits to a claim of personal authorship, the higher the disclosure stakes. This is the same gradient ghostwriting has always lived on; AI just put far more people on it.

Practical positions while the dust settles

You can't wait for the rules, because the social verdicts are already being handed out. A few positions that hold up:

Whatever policies eventually arrive will mostly ratify what audiences have already decided. The verdict taking shape is older than the technology: say what you mean, mean what you sign your name to, and the tools you used to say it are your business.