AI Content Disclosure

Every image and video produced by VIVID is generated by artificial intelligence. People, voices, products and scenes in Outputs are synthetic unless you supplied a real reference, and even then the result is an AI-generated depiction, not a recording of real events.

1. AI models we use

VIVID is an independent product. It is not made, endorsed or sponsored by the developers of the models below; we access them through our own infrastructure and add our own safety checks.

• Text to image: Qwen-Image (Alibaba Cloud, Qwen team).
• Image to image (editing an uploaded image): Qwen-Image-Edit (Alibaba Cloud, Qwen team).
• Text to video, first/last-frame video and reference video: MiniMax H3 (MiniMax).
• Prompt enhancement and content-safety checks: third-party large language models from our AI providers. They only rewrite or review text and images; they don't generate the final Outputs.

The model behind each feature is also shown in the workspace when you pick a task. If we change models, we update this page.

2. How we mark AI content

In line with Article 50(2) of the EU Artificial Intelligence Act, every Output is marked in a machine-readable format as AI-generated: we embed AI-generation provenance metadata (an XMP "AIGC" label with a unique production ID) into each generated image and video file. Please do not strip or alter this metadata.

3. Your disclosure obligations when you publish

If you publish or distribute an Output, you are responsible for disclosing that it is AI-generated where required. In particular:

• Deepfakes (EU AI Act Art. 50(4)). If an Output depicts realistic people, objects, places or events that could falsely appear authentic, you must clearly disclose that it is artificially generated.
• Advertising platforms. Meta, TikTok, YouTube/Google and other platforms require you to use their AI-generated / "synthetic content" labels for realistic AI content. Follow each platform's current policy.
• Endorsements and testimonials (US FTC; EU/UK consumer law). AI-generated "customers" must not be presented as real consumers, and AI content must not be used as fake reviews or testimonials. If an ad features an AI-generated person speaking about a product, make clear that the person is not a real customer — for example with an on-screen label such as "AI-generated. Not a real customer."
• Material connections. If you pay real creators or use a real person's likeness with permission, disclose that relationship as required by the FTC Endorsement Guides.

4. Suggested labels

• "AI-generated content"
• "This video was created with AI. The person shown is not a real customer."
• "Dramatization created with AI."

5. Limitations

AI models can produce inaccurate, biased or unexpected results, including text, hands, logos or product details that are wrong. Always review Outputs before use, and do not rely on them as factual representations of a product's appearance or performance.

Questions? Contact billykinggym@126.com. See also our Terms of Service and Acceptable Use Policy.