AI Transparency & Disclosure

7ART is a generative AI product. Everything you make with it — images, video, music, speech, 3D models, lessons and written copy — is synthetic content produced by machine learning models, not a recording of real events, people or performances. This page explains what that means in practice, what we mark, and what you are responsible for when you publish what you make.

You are always interacting with AI

Every generator on 7ART is automated. Our in-app assistant is an AI assistant, not a human support agent, and it is labelled as one wherever it appears. No human reviews your prompt before a generation runs. If you need to reach a person, contact support@7art.ai — that inbox is staffed by people.

What our models produce

• Images and video are generated or transformed by third-party models. A person, place or object shown may not exist, and a scene depicting something that looks real did not happen.
• Voices and music are synthesised. A voice that sounds like a specific person is an imitation produced by a model, not that person speaking.
• Lessons, articles and other written output are generated and can contain errors. Check anything you intend to rely on, teach from, or publish.
• Likeness features — talking photos, lipsync, character consistency — animate an image you supply. Applied to a real person, the result is a synthetic depiction of them saying or doing something they did not.

How generated content is marked

Outputs carry provenance information recording that they were generated by 7ART, together with the model used and the generation settings. We are extending this to machine-readable provenance marking in line with the technical standards referenced by the EU AI Act. Marking travels with the file we deliver; we cannot control what happens to it after third-party re-encoding, and some platforms strip metadata on upload.

Your obligations when you publish

Rules about labelling AI content apply to whoever publishes it. Where you share output — especially anything depicting a real, identifiable person, or content on matters of public interest — you are responsible for disclosing that it is artificially generated, and for having the rights and consents you need. Several jurisdictions require this explicitly.

• Do not present synthetic depictions of real people as genuine.
• Do not use likeness features on someone without their consent.
• Do not remove or obscure provenance information from output you share.
• Follow the platform rules wherever you publish — many require an AI label of their own.

Our Community Guidelines prohibit non-consensual intimate imagery and impersonation outright, and the Terms set out the rest.

Models we use

7ART routes generations to third-party providers, currently including fal.ai and kie.ai, which host models from a range of vendors. The model behind any given output is recorded in that asset’s details in your library. The set of available models changes as providers release and retire them.

Limits and accuracy

Generative models make mistakes. They produce inaccurate text, misspelled labels inside images, anatomically wrong figures, and confident-sounding claims that are false. Output is provided as-is for you to review — do not treat it as factual, legal, medical or financial advice, and check educational material before teaching from it.

Questions

For anything about this page, how a specific output was generated, or to report content that misuses someone’s likeness, contact support@7art.ai.