AI on OnlyFans: what an agency can legally do with chatbots, virtual models and clones.

Yes, you can use AI on OnlyFans, and on Fanvue and MYM, but the rules are tighter than most agencies assume. AI-generated content has to be labelled, fans must not be misled into thinking a machine is the creator, and nobody's face or voice can be cloned without her written consent. This guide sets out the platform rules, then the law in the US and the UK, the EU AI Act, whose transparency rules have applied since August 2026 to anyone whose chatbot talks to European fans, and the AI clauses your agency contracts should contain.

Can you use AI on OnlyFans?

Three uses get lumped together as "AI OFM", and they carry different risks. An AI chatbot drafts or sends messages to fans and sells pay-per-view content. An AI model is a fully virtual persona with her own account. A clone is a model trained on a real creator's photos or voice, producing content she never shot. The platforms treat them differently.

On OnlyFans, the Terms of Service say two things that matter. AI-generated content must be "clearly and conspicuously" captioned as such. And if someone else helps a creator run her account, "this does not affect your legal responsibility". The detailed Acceptable Use Policy changes over time; have a virtual-model project checked against the current version before launch.

Fanvue is the most open: every AI creator shows an AI tag in the bio and goes through the same verification as human creators. Its Community Guidelines (updated 14 June 2026) require a clear and prominent disclosure on AI media, ban using AI to impersonate real people other than the account owner or to depict anyone who resembles a minor, and allow AI replicas of a real person only with that person's documented consent and identity verification. MYM, the French platform, allows AI for content and for managing messages, and requires written consent for any AI content showing a real person. Our comparison of OnlyFans, MYM, Fansly and Fanvue covers the rest of the differences.

Is AI chatting on OnlyFans legal in the US?

No federal statute bans AI chatting. The question is deception, and US law tackles it on three levels.

The FTC Act. Section 5 declares unlawful "unfair or deceptive acts or practices in or affecting commerce". A fan who pays for private messages is often paying for the feeling that the creator herself is on the other end. Telling him so when software is writing, or answering "yes, it's really me" when he asks, is the kind of misrepresentation that framework exists for. The FTC acts through investigations and orders; civil penalties generally follow a breach of an order or a trade regulation rule.

California's bot disclosure law. Business and Professions Code section 17941 makes it unlawful to use a bot to communicate with a person in California online "with the intent to mislead the other person about its artificial identity" in order to incentivise a sale. A bot is defined as an automated account where all or substantially all actions are not the result of a person. A clear, conspicuous disclosure removes the problem. Note the nuance: a tool that suggests replies a human chatter approves is less likely to be a "bot" than one that sends on its own.

Companion chatbot rules. California's SB 243, chaptered in October 2025, requires operators of companion chatbot platforms to give a clear notice that the chatbot is artificial and not human where a reasonable person could be misled, with a private right of action for the greater of actual damages or $1,000 per violation. Whether an OnlyFans chatting setup counts as a companion chatbot platform is untested; a persona built to sustain an ongoing relationship with a fan is close enough to take seriously.

Our guide on chatters as employees or contractors covers the employment side of the same teams.

And in the UK?

There is no UK statute requiring AI disclosure as such. Consumer law does the work. Under the Digital Markets, Competition and Consumers Act 2024, a commercial practice is a misleading action if it provides false or misleading information relating to a product, a trader or any other matter relevant to a transactional decision, or if its overall presentation is likely to deceive the average consumer. That regime has been in force since 6 April 2025, and the CMA can now impose penalties itself, up to £300,000 or, if higher, 10% of turnover.

For a UK agency, the practical line is the same as in the US: disclose in the bio that messages are managed by a team using AI tools, never deny it when asked, and never sell AI output as something the creator personally recorded.

Why the EU AI Act matters even outside the EU

Since 2 August 2026, Article 50 of the EU AI Act (Regulation (EU) 2024/1689) has required that people be informed when they interact directly with an AI system, unless it is obvious to a reasonably well-informed person. It also requires those who publish deepfakes, AI image, audio or video content resembling existing persons that could pass as authentic, to disclose that the content is artificial. The Digital Omnibus, Regulation (EU) 2026/1744, postponed the high-risk rules but left Article 50 on schedule; only the machine-readable watermarking duty for tools already on the market was pushed to 2 December 2026.

The Act reaches providers and deployers established outside the EU where the output of the AI system is used in the Union. A London or Miami agency whose chatbot talks to fans in Paris or Madrid is inside the scope. The disclosure duty in Article 50(1) falls on the provider, the business that develops a system and puts it into service under its own name; an agency using an off-the-shelf chatting tool is usually a deployer, but one that builds its own bot and runs it under its brand can become a provider. Fines reach EUR 15 million or 3% of worldwide turnover.

Cloning a creator's face or voice: whose consent?

A clone is built from a creator's photos, videos or voice recordings. Without her written consent, it is a legal problem on both sides of the Atlantic.

• US. Most states protect a right of publicity. In California, Civil Code section 3344 makes anyone who knowingly uses another person's name, voice, photograph or likeness to advertise or sell, without prior consent, liable for at least $750 or actual damages, plus profits and attorney's fees.
• UK. English law has no general image right, but since 6 February 2026 it has been a criminal offence to intentionally create a purported intimate image of another person without their consent (Sexual Offences Act 2003, section 66E, inserted by the Data (Use and Access) Act 2025). Recordings of a face or voice are also personal data under UK GDPR.
• EU. From 2 December 2026, the AI Act prohibits using an AI system to generate realistic intimate or sexually explicit images, video or audio of an identifiable person without her explicit consent.

Consent that holds up is a separate signed document: what is cloned (face, body, voice), for which content and platforms, within which limits (the same ones she sets for real content), for how long, and what happens when she withdraws. If a creator finds a clone she never authorised, she should keep dated evidence, report it to the platform, and use StopNCII.org to block intimate images on partner platforms. In the UK, the Revenge Porn Helpline (0345 6000 459) supports adults dealing with synthetic sexual content. Our guide to leaked content explains the takedown steps.

The AI clauses your agency contract needs

Most management agreements signed before 2025 say nothing about AI. Add a schedule covering at least:

1. Permitted uses, listed by tool and platform; anything not listed is excluded.
2. Clone consent as a separate, signed, revocable document.
3. Prior approval of generated content by the creator, or at least samples of each type.
4. Disclosure: compliance with platform labelling rules and Article 50, and no denial to a fan who asks.
5. Chatbot rules: banned topics (meetings, personal data, promises she will not keep), human review of sales messages, escalation to a person when a fan seems vulnerable.
6. Ownership: training data and the trained model stay tied to the creator; the agency has a licence limited to the contract, with no reuse or resale.
7. Termination: deletion of the model and data within a fixed period, with written confirmation.
8. Liability and revenue: the agency indemnifies the creator for non-compliant use, and her share of clone revenue is written down.

These sit alongside the points in our guide to the 12 key clauses of an agency contract.

For creators: what to insist on

Ask for a written list of every AI tool used on your account, messages included. Do not hand over photo sets or voice recordings for training without a signed document describing their use. Set the same limits for generated content as for real content: what you would not shoot, you can refuse to have generated. Require deletion of any model when the contract ends. And reread your current contract: a broad grant of "image rights in all media, in perpetuity" can be used to justify a clone, and it can be renegotiated.

This guide is general information, not legal advice for your situation. We work with agencies and creators on contracts and cross-border compliance; our agencies page explains how.

Your contactOlivier MaciejewskiFounder of OFM Legal · Former international business lawyer (HEC Paris, Arendt, Clifford Chance) · Official MYM partnerHis background Updated September 11, 2026

This article is general information and does not replace advice tailored to your situation. OFM Legal is not a law firm: court proceedings are handled by our partner lawyers.

Our partners

See all our partners

Read next

Agencies & managersOFM courses: how to chooseRead the article Agencies & managersBecoming an OFM managerRead the article Agencies & managersOnlyFans management contract, agency sideRead the article