AI Transparency Notices
What is an AI transparency notice?
An AI transparency notice is a short, clear message that tells users when they are interacting with an AI system.
- A user-facing explanation.
- Written in plain language.
- Focused on trust, clarity, and informed use.
- A legal disclaimer.
- A privacy policy.
- A technical explanation.
We need transparency notices because:
- Some laws (e.g. the EU AI Act) require users to be informed when AI is involved.
- Users should understand when outputs are AI-generated or AI-supported.
- Transparency helps avoid confusion, over-reliance, and loss of trust.
When is a transparency notice required?
At Maersk, application owners complete an AI risk assessment for each AI use case.
- If the assessment result is Limited-Risk AI, a transparency notice must be shown to users
- Designers are involved after the risk assessment is completed.
- While designers focus on how and where it appears in the UI, the wording is pre-approved.
Pre-approved transparency notice copy
The texts below are pre-approved baseline copies. They are meant to give you a safe starting point. You may adapt the wording to better fit:
- Your application’s tone of voice.
- The specific user journey.
- The UI constraints.
As long as the core meaning is preserved. If you plan strong modifications, are unsure about suitability, or the use case feels borderline reach out to the Data & AI Ethics team: dataandaiethics@maersk.com
Pre-approved transparency notice examples
1. Internal use - AI assistants / chatbots
“This is an AI-assisted system. Outputs may be inaccurate—always review before acting Not for autonomous decisions.”
“This content is AI-generated and may be incomplete or inaccurate.”
2. External (Customer Experience) - AI system interacting with external users
“This response was generated with the help of an AI system. If you’d prefer a human agent, contact us at [email].”
“This is an automated response designed to assist you promptly. If you need further assistance, please reply to this message.”
3. Employee chat - HR assistant
“This assistant uses AI to answer HR-policy questions. Responses are advisory and linked to your corporate account.”
Please note:
If the chat is anonymous, “linked to your corporate account” should be removed.
4. Learning & Development tools
“This learning platform uses AI to personalize your training experience. Interactive content may be generated or adapted by AI.”
Where should the notice appear?
Important!
The following is extremly important and should be seen as the most important design rule to follow.
- Place the notice where the user interacts with the AI:
- Near generated outputs.
- Next to action buttons (e.g. “Generate”, “Suggest”, “Analyze”).
- In line with the AI feature itself.
- Ensure visibility at the moment of interaction.
- Make it readable without extra clicks whenever possible.
- Don’t hide it in:
- Settings.
- Help pages.
- FAQs.
- Terms & Conditions.
- Don’t show it only once if AI use is ongoing.
- Don’t place it far away from the AI-driven feature.
When in doubt
If you are unsure about:
- Whether a notice is required.
- How much you can adapt the wording.
- Where or how prominent it should be.
Reach out to the Data & AI Ethics team:. dataandaiethics@maersk.com