A genuinely tricky one, and something nobody's fully agreed on yet. If AI helped write the email, the proposal, the post — do you owe anyone a heads-up? Here's how I think about it, without tying myself in knots.
🤝 The test: would they feel misled?
The simplest gauge I've found. If the other person found out exactly how you'd used it, would they feel deceived? If yes, say something. If no, you're almost certainly fine. Using it to tidy your grammar isn't anyone's business. Sending an AI-written sympathy note as if you'd poured your heart into it — that's a different thing entirely.
✅ Where disclosure barely matters
Most everyday business use is invisible and completely fine: drafting, summarising, brainstorming, fixing spelling, first drafts you then rework. Nobody labels their spellchecker or their calculator. AI as a behind-the-scenes assistant sits in the same bucket. You did the work; it helped you do it faster.
⚠️ Where it really does matter
When the human-ness is the point. A personal message that's meant to come from you. Creative or written work someone is paying you to produce as your own. Anything where authenticity, authorship, or "a real person actually looked at this" is the value being sold. Passing AI work off as fully your own there crosses a line — and getting caught costs you trust you can't easily buy back.
🧾 The accountability bit
Whether or not you disclose, you own the output. "The AI said so" is not a defence to a client, a customer, or anyone else. If you send it, you're standing behind it — which is exactly why you read it properly first. Disclosure is about honesty; accountability isn't optional either way.
🧭 A simple working rule
Don't hide it, don't over-announce it. If someone asks, be straight with them. For anything where a person is paying for your personal touch or judgement, make sure your touch and judgement are genuinely in there — then it's honest by default. Transparency is cheap; trust is expensive. Spend accordingly.