AI Interaction Notice and Synthetic Content Labeling

How to use this document

• Use it to tell people they are interacting with an AI system and to label the AI-generated or AI-manipulated content (images, audio, video and text) that the organization publishes.
• Suggested owner: the system owner of each channel (with customer service for interaction notices and marketing or communications for content labeling); Legal reviews both. The general transparency criteria and the explanation circuit are in the "AI Transparency and User Explanation Policy".
• Before adopting it, decide: which channels use AI (chat, voice, email, social media), whether the organization acts as a provider (its own system or one put into service under its brand) or as a deployer (a third-party tool), and whether it publishes synthetic content showing real people, places or events.
• Complete the bracketed fields: assistant name, keyword to reach a person, channels, owners and links.
• Use each text in section 4 only in the channel it is designed for, and adapt the tone without removing the core message: "this is AI, not a person."
• Keep in "Country variants" only the jurisdictions where you operate; application dates appear in the regulatory references at the end.
• Ask internal legal or an outside advisor to review the final texts, especially in health care, financial services, minors and political campaigns.
• Approve the document in the [AI COMMITTEE], publish the notices in each channel and record each deployment in the table in section 8.

AI Interaction Notice and Synthetic Content Labeling

1. Purpose and scope

[ORGANIZATION NAME] adopts this notice and procedure so that anyone interacting with one of the organization's AI systems knows they are not dealing with a person, and so that AI-generated or AI-manipulated content published by the organization is identified as such.

It applies to:

• conversational assistants (chatbots), agents and voice assistants that serve customers, users, candidates or the public;
• calls, messages and emails generated or sent by AI systems;
• images, audio, video and text generated or modified with AI that the organization publishes or distributes;
• [OPTIONAL] emotion recognition or biometric categorization systems the organization uses with people.

It does not apply to AI tools used only internally that do not interact with third parties or produce published content.

2. Definitions

• Interaction notice: a message telling a person that they are interacting with an AI system.
• Synthetic content: an image, audio, video or text generated or substantially modified by an AI system.
• Deepfake: synthetic content that resembles real people, objects, places, entities or events and could falsely appear authentic.
• Technical marking: a machine-readable mark (metadata, watermark, content credentials) that allows content to be detected as synthetic.
• Visible label: text or an icon that a person can see or hear and that indicates the content is synthetic.