AI & Content Disclosure Policy

1. Disclosure is on by default

1.1When we publish content to a connected account, AI disclosure is turned on. It is not an option that can be switched off. You can adjust the wording and where it appears; you cannot remove the fact of disclosure, and no publishing route we offer, including the API and MCP interface, omits it.

1.2A platform that detects undisclosed synthetic media may penalise your account, not ours. We turn disclosure on to protect you as well as your audience.

2. How disclosure is carried

2.1Metadata does not survive most platforms, so disclosure is carried in more than one way:

• Platform labels: we set the platform's own AI-content label or toggle on every publish where the platform provides one, and record that it was set.
• Caption wording: where a platform needs words rather than a label, we add disclosure text, using your workspace's disclosure wording.
• On-screen labelling: for cases that call for it, such as synthetic likenesses of real people and virtual-influencer content, the label appears on the content itself and cannot be turned off.

2.2If a platform offers no label, we record that no platform label was available, so a later question about what was disclosed can be answered.

3. Content credentials and provenance records

3.1Where we have signed an asset with a content credential (an industry-standard, cryptographically signed manifest), it states that the content is AI-generated, when it was made, and who issued the credential. If you choose attribution, it can name your workspace. It does not contain the prompt, the AI model or provider used, or your reference images: we disclose the nature of the content, not your inputs or our supply chain. Credentials are being rolled out, so an asset may be delivered without one; the labels above apply either way.

3.2Many platforms strip or rewrite metadata when you upload, so a credential may not survive distribution. Do not rely on it as the only disclosure.

3.3Independently of credentials, we keep a generation record for each piece: the generation history, the prompt version, and the references that shaped it. It is evidence of how the piece was made, and it is kept under legal hold if a claim is made. We do not alter output to hide that it is AI-generated.

4. Checks we run, and what they are not

4.1We run automated and human checks before publishing: content moderation, flags for third-party marks and characters, likeness and voice consent, and checks of claims in advertising. They reduce risk. They do not establish that content is original, free of third-party rights, lawful in your market, or fit to publish, and we make no such statement. See the Terms on intellectual property and the absence of an indemnity.

5. Your responsibilities

5.1You remain responsible for complying with the laws and platform rules that apply to your content and audience, including AI-transparency and deepfake-disclosure rules such as those in the EU, and advertising, endorsement and consumer-protection rules. Do not remove or hide a label or caption we add. Do not present AI-generated content as filmed footage of real events or of a real person.

5.2If you share Outputs outside the Service, for example by downloading and posting manually, add disclosure yourself where the law or the platform requires it.

6. Questions and reports

6.1Write to [email protected]. To report undisclosed or misleading synthetic content made through the Service, use the report form or [email protected].