AI tools, chatters and your likeness What an agency may do with your face, voice and chats

Platforms have different rules on AI replies and AI versions of you. Here is what each says, what the EU requires, and what to put in your contract.

By the Crewvale editorial team

Three different things get called “AI” in creator work. Software that writes or suggests replies to your fans. An AI version of you, built from your photos, voice or chats. And fully synthetic creators who don’t exist. Each has different platform rules and different legal questions. If an agency uses any of them in your name, you want it in your contract before you start.

In short

Platform rules differ. Fansly bans photorealistic AI content and AI that talks to fans without human oversight. Fanvue allows AI content with clear disclosure. OnlyFans allows AI content only if it clearly features the verified creator and users can clearly tell it is AI generated.
Nobody should make an AI version of your face, voice or style without your written consent. Your contract should say whether they may and for how long.
The EU AI Act sets transparency duties for AI that talks to people and for synthetic content. Check how they apply to you with a lawyer.
You stay answerable to the platform for what is posted on your account.

This is general information, not legal advice.

What do agencies use AI for?

You may meet it in four places:

Reply suggestions. Software drafts a message that a chatter edits or sends.
Automatic replies. A bot answers fans with no person in the loop.
A model of you. A system trained on your photos, voice or past chats, so that it can write like you or produce new images of you.
Synthetic content. Images or video that look like you, or like someone who doesn’t exist.

The first is the mildest. The last two are the ones to put in writing.

What do the platforms say?

Rules change. Read the current page before you act. This is what each platform said on 25 October 2026.

Platform	AI-generated content	AI in messages	Real people
OnlyFans	Allowed only if it clearly features the verified creator, a user can clearly tell it is AI generated (for example #ai or #AIGenerated), and it follows the platform’s terms.	Check OnlyFans’ current terms before you use AI-assisted replies.	You are legally responsible for all content you upload, even if someone else helps run your account.
Fansly	“Photorealistic AI-generated content is not allowed.” Labelling it does not change that. Non-photorealistic virtual personas may apply under the owner’s real legal identity.	Its terms prohibit AI that directly interacts with users to generate personalised responses. The help article describes this as AI without human oversight.	Deepfakes and face-swapped content are covered. Creators must never impersonate anyone without clear, explicit consent.
Fanvue	Allowed with clear disclosure in your bio, a caption or a watermark. Fully AI accounts must never tell fans they are a real person.	The same rules apply “including in messages”.	AI media of a real person other than the account owner is not allowed under the policy section. The help article describes consent routes for AI personas based on a real person. The policy is stricter, so follow it.

Two points follow. If your agency uses AI to write to your fans, check which platform you’re on. And “human oversight” and “disclosure” are not the same thing. Fansly cares about the first, Fanvue and OnlyFans about the second.

Do fans have to be told that AI is writing?

Platform rules decide this first. Fanvue requires clear disclosure of AI content, and bars fully AI accounts from claiming to be real people. OnlyFans allows AI content only if it clearly features the verified creator and users can clearly tell it is AI generated.

The law is catching up. Article 50 of the EU AI Act sets transparency duties. Under it, providers must make sure AI systems that interact directly with people tell them it is an AI system, unless that is obvious from context. Providers of systems that generate synthetic images, audio, video or text must mark the output in a machine-readable format. Deployers of systems that make deep fakes must disclose that the content is artificially generated or manipulated. A lighter duty applies to evidently artistic, creative, satirical or fictional works.

Dates and exact wording can change, so check the official EUR-Lex text before you rely on them.

What about an AI version of you?

Making an AI version of a real person from their photos or voice without their consent is a risk for everyone involved. Platforms are clear. Fansly says creators must never impersonate anyone without clear, explicit consent. Fanvue’s policy section says AI media must not be of a real person other than the account owner.

For your own contract, treat it as a licence question. Ask:

May the agency create or train anything from your photos, voice or chats? If yes, for what purpose?
Who owns the result?
Can it be used after you leave? On other accounts? For other creators?
Must the agency delete it, and prove it did, when the contract ends?

A clause that gives the agency “all rights in your likeness, in any medium, in perpetuity” covers all of this. Strike it or narrow it. See the contract terms guide.

What happens to your chats?

Messages between you and fans are material for a model. A tool that learns from them could, in principle, learn to write like you.

Ask your agency in writing:

Which tools touch your chats, and who runs them?
Are your chats used to train or improve a model?
Where are the chats stored, and are they deleted when you leave?

Data-protection law may give you rights over how your chats are used, so check the rules where you live.

Can content be marked as AI or as real?

There is a technical standard for this. An industry group publishes Content Credentials, an open standard to show where digital content came from and how it was changed, which it describes as working “like a nutrition label for digital content.” The EU text above also points to machine-readable marking of synthetic output.

You can ask your agency whether its tools add such information to files. It is a useful question, and a vague answer tells you something.

What should your contract say?

Add a short AI clause. Ask for these points in writing:

No AI version of you without your separate written consent, per use.
No fully automated replies in your name. A person reads and approves messages.
Disclosure wherever the platform or the law requires it, and the agency bears the cost of a breach it causes.
No training on your photos, voice or chats, and deletion of any such data when the contract ends.
Takedown. If anything made from your likeness is published, the agency removes it on request within a stated time.
Liability. If a platform bans or restricts your account because of the agency’s AI use, the agency says in writing who bears the loss.

The last point matters. OnlyFans says you stay legally responsible for what is uploaded to your account. So a ban caused by your agency’s tools is a loss you carry unless the contract says otherwise.

What if your likeness is used without your consent?

Gather evidence first: screenshots, links, dates. Report it to the platform, using its impersonation or privacy route. If the agency did it, write to them and ask for removal and a copy of what they hold. A lawyer can then advise on a takedown and on claims. Nobody should apologise for asking.

How does Crewvale fit in?

Crewvale is a free marketplace for the OnlyFans and webcam industry: creators and cam models compare OnlyFans and webcam agencies on their published terms, agencies meet creators looking for representation, and agencies hire staff on its jobs board. Crewvale isn’t an agency and doesn’t handle payments.

Read an agency’s published terms next to this list. Use the contract checklist and, if an agency misuses your likeness, report a concern.

Sources

All read on 25 October 2026.

Last reviewed: 25 October 2026.