What should a company AI disclosure policy actually say, in plain terms?
I lead trust and safety at a mid-sized marketplace and I have been asked to draft an AI disclosure policy for the company. Not the legal terms of service, an internal and external statement of how we use AI and what we tell people about it.
Cite
Global 100 Forum, "What should a company AI disclosure policy actually say, in plain terms?", https://global100.org/forum/q/what-should-a-company-ai-disclosure-policy-actually-say-in-plain-terms/, accessed 2026-11-21.Every template I find is either so vague it says nothing ("we use AI responsibly") or so detailed it reads like a regulation nobody could follow. I would like to write something a seller, a buyer and a new employee could each read in five minutes and understand.
Specifically: what has to be in it? Who is covered, employees, contractors, vendors? What counts as AI use that needs disclosing, and what is too trivial, like spellcheck? What happens when someone gets it wrong? I would value hearing from people who have written one, been subject to one, or enforced one.
Write something first.
Give people something to work with: at least 30 words on what happened and what you tried.
That is too long. Keep it under 6,000 characters.
Write the question as the title, 15 to 140 characters, no links.
Pick a category.
Add a name (2 to 40 characters, no links).
That email address does not look right.
That was quick. Read the thread, then try again.
The form expired. Reload the page and post again.
Something went wrong with the form. Reload and try again.
Please complete the check and post again.
Limit reached for now. Try again later.
This thread is closed to new replies.
This thread no longer accepts replies.
Something went wrong with the form. Reload and try again.
Post a reply or question first (name and email), then this browser can vote, edit and accept answers.
You cannot vote on your own post.
Only the author (within 30 days) or the forum team can do that.
That email already has an account. Use your sign-in link instead.
New accounts are paused for the moment. Try again later.
That was already posted.
2 answers
- Tobias K.Member ·
Procedurally, I would structure it in five parts and keep each to a paragraph. Scope: who and what is covered. Definitions: what you mean by AI-generated and AI-assisted. Duty: when disclosure is required and to whom. Process: how someone reports use and who reviews it. Consequences: what follows a good-faith mistake versus a deliberate omission. Most policies I see fail at the fourth and fifth, which is where people actually need guidance.
- Lena B.Member ·
Freelancer view, because contractors get forgotten. Say plainly what you expect from outside people. Do I need to tell you if I use an assistant for a first draft? For research? For editing? If the policy is silent, I will guess, and I will guess in my own favour. If you want to be told, say so and say how, in one line on the invoice or the delivery email.
Also say what you will not do. No detector scores as the basis for dropping someone. That sentence earns more trust than any principle.