AI governance
AI disclosure: what to tell people, when to tell them and how to set a policy
A plain guide for compliance, legal and operations leads deciding when and how to say that AI was used. It separates the different meanings of the term, summarizes one source of rules, and sets out a simple policy to agree with counsel.
AI disclosure is telling people that artificial intelligence was involved in something they are dealing with, such as a chat, a document, an image or a decision. The term is used in several ways: telling customers they are talking to an AI system, labelling content that AI generated, telling staff which AI tools they may use, and telling an applicant that AI played a part in a decision. Rules differ by country, state and industry, and some are still taking effect, so a business should confirm what applies to it with counsel. A sound starting point is a short written policy that lists each place AI touches a person, who decides what to disclose, and the wording to use.
plain answer
What is AI disclosure?
AI disclosure means being open that AI was used. The reason is trust: a person who knows a chat answer, a report or a decision involved AI can judge it differently, ask for a person, or check it. A disclosure can be a line of text, a label on content, a notice in a terms page, or a statement in a formal filing.
Because the phrase covers several situations, it helps to separate them before writing any policy.
use cases
When should a business disclose AI use?
These are the situations that come up most often. Which ones are required, and in what form, depends on your location and industry.
| Situation | What disclosure usually means | Questions to settle |
|---|---|---|
| Customer chat and voice | Telling people they are dealing with an AI system and not a person. | Where the notice appears, and how a person can ask for a human. |
| Generated content | Labelling text, images, audio or video that AI created or changed. | Which content needs a label, and who reviews it before it is published. |
| Decisions about people | Telling an applicant, patient or employee that AI supported a decision about them. | What reasons you can give, and how a person can ask for review. |
| Staff use of AI tools | Telling employees which tools are approved and what data may go into them. | Who approves tools, and what happens when staff use an unapproved one. |
| Public statements and filings | Describing your use of AI accurately to investors, regulators or the press. | Who checks that claims about AI match what the systems actually do. |
governance
What does the EU AI Act say about AI disclosure?
The EU AI Act includes a transparency article, Article 50. As summarized from the European Commission's AI Act Service Desk (read 2026-11-11), if the Act applies to your business, it covers several cases. Providers must design AI systems that interact directly with people so that those people are informed they are interacting with an AI system, unless this is obvious to a reasonably well-informed person. Providers of systems that generate synthetic audio, image, video or text must mark outputs in a machine-readable format so they can be detected as artificially generated or manipulated. Deployers of emotion recognition systems must inform the people exposed to them. Deployers of deep fakes must disclose that the content is artificially generated or manipulated, and deployers of AI-generated text published to inform the public on matters of public interest must disclose that too, unless the text has been through human review or editorial control. The information must be given clearly, at the latest at first interaction or exposure. Each duty has exceptions.
This is a summary of one source, not legal advice, and the page is not the Official Journal. Check the official text on EUR-Lex, and our EU AI Act page for dates. Other rules, such as US state laws, sector regulators and consumer protection law, can also apply and are outside this page.
how to
How do you set an AI disclosure policy?
A short written policy is enough to start. Agree it with legal counsel.
1. List every place AI touches a person
Include chat, email drafts, generated content, scoring and decision support, and internal tools used on customer data.
2. Check what rules apply to each
Ask counsel which laws and sector rules cover each use in the places you operate.
3. Write the wording
Use plain sentences that say what AI did and how to reach a person. Avoid vague labels such as AI-powered.
4. Name an owner
One person approves disclosures, keeps the list current and answers questions.
5. Test and record
Check that notices actually appear where they should, and keep a record of what was shown and when.
6. Review on a schedule
Rules and your own AI uses change. Revisit the policy at least when you add a new use.
kriv fit
Where does Kriv AI fit?
Kriv AI helps regulated teams put AI to work with controls designed in from the start. For disclosure, that usually means building the list of AI uses, drafting the policy and wording for counsel to review, and training the people who will apply it. Kriv AI does not give legal advice. We scope work after a short discovery call.
Straight answers
Frequently asked questions about AI disclosure: what to tell people, when to tell them and how to set a policy
What does AI disclosure mean?
It means telling people that AI was involved in something they deal with, such as a chat, a piece of content or a decision about them.
Do I have to tell customers they are talking to an AI?
In some places the law requires it. The EU AI Act, for example, has a rule for systems that interact directly with people. Rules vary, so confirm what applies to you with counsel. Many businesses disclose by choice.
Do I need to label AI-generated content?
Some rules require labelling of certain AI-generated content, and requirements differ by place and type of content. Check with counsel, and decide your own standard for content that could mislead.
What should an AI disclosure say?
Say what AI did and how a person can be reached, in plain words. Avoid vague labels that do not tell the reader anything.
Who should own AI disclosure in a company?
One named person, usually in compliance, legal or operations, who keeps the list of AI uses and approves the wording.
Talk to the team that would do the work
Bring your requirements to a working session with the person who'll actually deliver.
Book a Discovery Call