Fanvue allows AI-generated creators, but its own rules attach conditions to every synthetic image, video and audio clip: disclose the AI clearly and prominently, never depict a real person other than the account owner without documented consent and identity verification, never produce anyone who resembles a minor, and keep a verified human behind the account who has passed Fanvue's identity and liveness checks.
This checklist is built only from Fanvue's published documents: its Community Guidelines, Acceptable Use Policy, Creator Terms, content moderation policy and Help Centre articles. It covers Fanvue alone. Other platforms draw the line elsewhere; our guide to OnlyFans AI and deepfake rules covers that platform, and the Fansly creator explainer notes that Fansly does not allow photorealistic AI content at all.
Where Fanvue writes its AI rules
The controlling text is section D, AI-Generated Media, of the Fanvue Community Guidelines, which define AI-generated media as anything made with artificial intelligence, including deepfakes, faceswaps and synthetic audio, video and imagery. The Acceptable Use Policy repeats the disclosure duty under content authenticity, sets identity rules for face coverings and defines the public media that must stay safe for a general audience.
Three Help Centre articles add the practical detail: Is AI content allowed on Fanvue?, What is an AI-Generated model? and Passing KYC and creating multiple accounts as an AI Creator.
Watch for version drift between them. The first Help Centre article reproduces the policy as it stood on an earlier date and its wording does not match the current guidelines word for word; for example, it lists an earnings deduction among possible consequences and describes how the moderation test is staffed, details the guidelines phrase differently. That article itself tells readers to read the Community Guidelines before signing up. A sensible working rule is to treat legal.fanvue.com as the authority and, where the help text is stricter or more specific, follow that as well.
AI model or real model: which category you are in
Fanvue separates fully synthetic characters from synthetic edits of real people, and the obligations change with the category. The AI-Generated model article says that deepfakes and face-swaps are not AI-Generated Models, and that if a model's body belongs to a real person, even with an AI face, the model is classed as a real model and must be verified as one.
| Scenario | How Fanvue treats it | What the rules then require |
|---|---|---|
| A fully synthetic character, face and body generated | An AI-Generated Model | Disclose the AI in your bio, never tell fans the character is a real person, and pass KYC as the account owner. |
| An AI face placed on a real person's body | A real model | The person whose body appears must be verified and must have consented, through one of the routes in the KYC article. |
| A face-swap or deepfake of an identifiable person | Deepfaked or AI-replicated media | Documented consent and identity verification of the original individual before anything is published. |
| A lookalike of a celebrity or someone you know, made without their involvement | Impersonation of a real individual other than the account owner | Not permitted; the guidelines bar misleading or impersonating synthetic media of real people. |
| An AI character wearing a mask or face covering | Acceptable for AI media | Disclose it as AI and do not let it pass as a real person; real creators cannot use coverings that block identity checks. |
| Synthetic voice notes or audio | AI-generated media under the definition | Disclose the synthetic voice in the message or caption that carries it. |
| A real creator's own photos run through AI tools | Not spelled out; the definition covers media generated using AI without saying where editing ends | Conservative default: if AI generated or materially changed what viewers see, disclose it. |
The last row is the grey area most likely to trip up a real creator. Light retouching and an AI-built body or background are different things, but the documents do not draw that boundary precisely, so disclosure is the cheaper mistake. If a piece of content sits on the line, ask Fanvue support before publishing and keep their reply.
Disclosure: what clear and prominent looks like
The guidelines require every piece of AI-generated media to carry a clear and prominent disclosure and give a watermark, a caption, an accompanying message or a bio statement as examples. The Help Centre version adds that you must say the media is not real and was generated using AI, so other members know it is not authentic, and that failing to disclose properly can lead to removal or suspension. The AI-Generated model article goes further for persona accounts: disclose in your bio that the model is AI, and do not tell fans you are a real person in posts or in chats.
| Surface | What Fanvue's documents say | A practical way to comply |
|---|---|---|
| Profile bio | State that the model is AI | Put a plain statement in the opening line, not after a run of emojis or hashtags. |
| Profile picture, banner, intro video and public bio text | Public media must be safe for a general audience and AI imagery there needs clear disclosure | Keep these SFW and add a visible AI marker to the banner as well as the bio statement. |
| Feed and paid posts | All AI media needs a clear and prominent disclosure, and a bio statement is one accepted form | Use a caption marker or watermark on every item so it still reads as AI when seen out of context. |
| Direct messages | An accompanying message counts as disclosure; never claim to be a real person in chats | Build the disclosure into message templates and brief anyone who writes messages for the account. |
| Audio and voice notes | Synthetic audio falls inside the AI-generated media definition | Name the voice as synthetic in the text that accompanies it. |
The public media rules come from the Acceptable Use Policy and section F of the guidelines, which define public media as the profile picture, banner, intro video and public bio text that people see before signing in or subscribing. Those areas must not include nudity, implied nudity, see-through clothing, hyper-sexualised posing or explicit language, and AI imagery there must be disclosed. Treat disclosure as part of the asset, not a caption you remember afterwards.
Likeness, consent and age: the hard limits
Section D of the guidelines sets limits that disclosure cannot fix. Synthetic media must not be misleading or harmful or be used to impersonate real individuals other than the account owner. Deepfaked or AI-replicated media needs the documented consent and identity verification of the original person before it is published. AI-generated individuals must not resemble minors, and prompt text or disclaimers are not enough to overcome a visual resemblance.
The Help Centre AI article explains why prompts are not proof: the model, its training data, prompt weighting and other variables all shape the output. In practice this means the review happens on the finished image or clip, never on the prompt that produced it. Reject anything with ambiguous age cues, and keep youthful styling, props and wording away from adult personas. The Acceptable Use Policy separately bans age-baiting language in public media, giving “just turned 18”, “barely legal” and “teen” as examples.
Rights questions run in the same direction. The guidelines say users may only upload content they have the legal right to post, as the rights-holder or under a valid licence, and the Help Centre article says AI media must not infringe copyright, that the uploader is responsible for having the right to use and share it, and that content taken from other websites or creators without explicit authorisation is not allowed. Before building a persona, check that your generation tool's terms let you use outputs commercially, and never train or prompt on another creator's photos. The documentation habits in our content rights release checklist transfer well to AI workflows that involve any real contributor.
Verification for the human behind the account
An AI persona does not remove the onboarding requirements. Fanvue's Creator Terms require every creator to submit valid identity documents and pass a liveness check, provide accurate tax and banking details, declare whether content is explicit and agree to the platform policies, and Fanvue reserves the right to ask for re-verification at any time. The KYC article adds that even when your face never appears in the content, you need a valid, in-date government-issued ID or passport and a selfie, and that this information is not shown on your profile.
Featuring any other real person, including a real body under an AI face, triggers a second set of steps. According to the KYC article, either that person opens their own Fanvue creator account, completes KYC and is tagged every time you share media featuring them, direct messages included, or you complete Fanvue's model release form for them and send it with their ID, a full-body photo and a headshot of them holding paper showing the current date and your username, with a new form for each piece of content that features a new individual. The Co-Authored Content Policy also requires you to keep their ID and signed, dated, fully informed written consent on file, and notes that earnings go only to the uploading account, with any split left as a private arrangement.
Once verified, you can open linked creator accounts from Settings without re-verifying, but the KYC article says you must have made a first withdrawal on the original account, that linking is unavailable while any account has an active warning, and that the account limit is 2 for Human Creators and 15 for AI Creators.
Expect detection as well as self-declaration. Fanvue's content moderation policy says uploads are first scanned by third-party AI-based moderation that flags, among other things, whether content has been generated by AI, with flagged items hidden and queued for human review, and that new creators receive a manual review before their first withdrawal.
What happens when content fails a check
AI media is assessed under what Fanvue calls the Reasonable Person's Test. The Help Centre AI article describes it as conducted by three or more members of the moderation team, says a failed test removes the content and warns that multiple failures may lead to permanent suspension.
The guidelines list content removal, temporary suspension or a permanent ban depending on severity. The Acceptable Use Policy goes further on money: Fanvue may revoke a creator balance in part or in full, prevent withdrawals and remove media in bulk from accounts with a substantial volume of violating content. For public media, a first offence takes the profile out of the Discover section until the media is fixed, and later offences hide the whole profile from public view, although neither step deletes content or affects earnings from existing subscribers.
Appeals run on two clocks. The content moderation policy gives 14 calendar days from an enforcement action to appeal through support, while the Creator Terms allow six months to dispute an earnings forfeiture or an account suspension or termination under the complaints policy. Diary the shorter one.
Pre-post compliance checklist for AI creators
Run the account section once at launch and again whenever Fanvue updates its policies. Run the asset section on every upload, including media sent in messages.
Account set-up
- The account owner has passed identity verification and the liveness check under their real legal identity.
- The bio says plainly, near the top, that the persona is AI-generated.
- Profile picture, banner, intro video and public bio text are SFW, carry an AI disclosure and contain no age-baiting words.
- You know which category applies: fully synthetic model, or real model because a real body, face or voice is involved.
- Every real contributor is either a verified Fanvue creator you tag on each share, or covered by a model release, ID and dated photos on file.
Every asset before upload
- A disclosure is on the asset itself or in its caption, not only in the bio.
- The finished output has been reviewed for any resemblance to a minor, and anything ambiguous has been discarded.
- The output does not resemble a celebrity or any identifiable person who has not given documented consent and been verified.
- You hold the rights to every input and output, and no other creator's content was used to make it.
- Intimate content carries Fanvue's adult-content label and nothing in it falls into the prohibited categories of the Community Guidelines.
Messages, team and records
- Message templates include the AI disclosure, and no one writing for the account claims the persona is a real person.
- Synthetic voice notes are labelled in the accompanying text.
- No one requests or accepts payment off the platform or promotes external payment links, which the Creator Terms treat as grounds for suspension and payout withholding.
- You keep a dated log of disclosures, source assets, releases and any support guidance on borderline content.
- You know the appeal windows and who on your team watches for enforcement notices.
Limits of this checklist
This is a reading of Fanvue's public documents as they stood on 1 October 2026, not legal advice and not an official Fanvue interpretation. The guidelines say Fanvue may amend its policies and that the latest version always lives at legal.fanvue.com, and the Help Centre can lag behind the legal pages. Re-check the sources linked above before launching a persona and whenever you receive a policy notice.
The Reasonable Person's Test is a human judgement, so following every item here lowers your risk without promising that a given post will pass. Laws on synthetic media, deepfakes and intimate imagery differ between countries and are changing quickly, and a platform rule does not change your exposure under local law; speak to a lawyer in your jurisdiction before building a business on a synthetic persona. If someone uses your own likeness without consent, the evidence approach in our impersonation reporting workflow is a useful starting point, alongside Fanvue's own reporting tools.