AI in Creator Work
What Should I Disclose About AI to My Subscribers?
The OnlyFans policy in plain language, the disclosure phrasing that retains trust, and the difference between assisted work and replacement work.
Two questions get asked the most by creators on this topic. Do I have to tell my fans about AI. And if I do, how do I say it without sounding like I am apologizing.
This piece answers both. Platform policy in clean language. The legal line between assisted work and replacement work. Three sample disclosure paragraphs at three transparency levels, with notes on which one actually earns trust. And a task-by-task table for what AI is doing in your account and what your fans deserve to know about it.
The OnlyFans policy in plain English
The platform has published guidance on AI-generated content. Most creators have not read it directly. The short version is two clauses.
One. AI-generated content that depicts a person must match the verified creator’s likeness. If your verified identity on the platform is Maria, AI-generated images on your account must look like Maria. AI-generated images of someone who does not look like you are a policy violation, and the account is at risk of removal.
Two. AI-generated content must be disclosed. The platform does not dictate the exact wording. A line in the post, in the caption, or in your account bio that identifies the content as AI-assisted or AI-generated meets the requirement. Posting AI-generated content without any disclosure is the violation.
Both clauses apply specifically to the visual content itself. The platform does not, as of 2026, require disclosure for AI used in scheduling, in caption writing, or in drafting fan messages that a human then edits and sends. Those uses are between you and your fans as a trust matter, not a platform policy matter.
Fanvue is a separate platform, with separate rules. Fanvue permits fully AI personas with no real-person backing. OnlyFans does not. Creators sometimes mix the two up and assume what works on one works on the other. It does not.
Assisted work versus replacement work
A useful distinction, because it determines what you need to disclose and what you do not.
Assisted work is AI handling tasks adjacent to your creator role, where every decision that reaches a fan still passes through a human. Scheduling. Caption variants. Draft text for fan replies that an operator edits before sending. Photo resizing for cross-platform posting. Pricing recommendations on PPV that a human approves. In assisted work, AI is a tool. You and your team are still the operators.
Replacement work is AI standing in for you on something fans pay you to be. Generating images that depict you, without you being in them. Voice cloning your responses to fans. Producing video that did not come from a shoot you sat for. In replacement work, AI is the deliverable. The fan is paying for you and receiving an AI substitute.
The platform policy applies to replacement work. Assisted work is not regulated by the policy because no fan-facing content is being synthesized. Reputational trust applies to both, in different ways. Replacement work requires the disclosure the platform requires. Assisted work requires the disclosure your fans expect, which is a different thing in a different register.
Sample disclosure language at three levels
Pick the one that matches how you actually run your page. A disclosure that does not match the reality is worse than no disclosure.
High transparency
“Hi, before you message me — I want to be upfront about how this page works. I’m a real person and you can verify that by asking for a custom video chat or a personalized voice note. I work with a small team that helps me run the page when I’m filming or off the platform. Scheduling, posting, and some of the fan replies are drafted by my team with help from AI tools. Every reply gets reviewed by a human before it sends. The content you see on my feed and in PPV is filmed by me. If you ever want to be sure you’re talking to me directly, just ask.”
Notes. This earns the most trust from the fans who think the most about this question. It also costs almost nothing to write, lives in a pinned post or a welcome message, and is a one-time investment. Fans who read this and stay are likely to spend more, not less, because they feel respected. The high-transparency version is also the safest from a platform-policy standpoint because it pre-discloses everything.
Medium transparency
“I work with a team that helps me run the page. Some scheduling and replies are assisted. Everything that matters is me. If you want a custom voice note or a video chat to verify, just ask.”
Notes. This is the most common phrasing in well-run accounts in 2026. It tells fans the truth at the level they actually care about, offers a verification path, and does not over-explain. The fans who want more detail will ask. The fans who do not, are satisfied with the headline.
Low transparency
“This page is run with the help of a team.”
Notes. This is the floor of acceptable disclosure. It does not commit a fan to any specific belief about who they are messaging, but it does not claim every reply is the creator’s own typing. The trust risk is moderate. If a fan finds out specifics that the low-transparency version did not cover, the surprise feels manageable rather than betrayed. The risk is that fans who would have preferred medium or high transparency do not get the information they wanted, and may churn after a few months of suspicion.
The bad version, the one that breaks trust, is the disclosure that claims something untrue. “Every message you receive is typed by me personally” on an account where draft tooling is in use, is the version that produces the Reddit threads that end creators’ incomes. Do not write that, and do not let any agency you work with write it for you.
Task-by-task: what AI is doing, what fans need to know
A reference table for the tasks that come up in a working account.
| Task | AI involvement | Disclose to fans? |
|---|---|---|
| Posting schedule | Yes, time-of-day optimization | Optional, low-risk |
| Caption writing | Hybrid, AI drafts, human picks | Optional, low-risk |
| Fan message drafting | Hybrid in most agencies | Recommended |
| Fan message sending | Human always | No disclosure needed |
| Content production (filming) | No, human always | No disclosure needed |
| Image variation, resizing | Yes, automated | Optional, low-risk |
| Voice notes | Human in well-run accounts | If AI voice is used, disclose |
| Video chats | Human always | No disclosure needed |
| PPV pricing | Hybrid, AI recommends, human sets | Optional, low-risk |
| AI-generated images of you | Possible, with policy constraints | Required by platform policy |
| AI-generated video of you | Possible, with policy constraints | Required by platform policy |
The pattern. Anything that touches a fan directly and is fully automated should be disclosed in some form. Anything that is human-final, with AI as an internal tool, is optional disclosure where the recommendation depends on how transparent your overall pattern is. Anything that synthesizes your likeness is required disclosure by the platform.
What disclosure does to your revenue
The fear underneath this entire question. Creators worry that telling fans about AI involvement will cost them money. The fear is reasonable. The data does not support it.
The accounts that disclose at a medium or high transparency level retain fans at rates indistinguishable from accounts that do not disclose at all, in the short term. In the long term, they retain fans better, because the disclosure pre-empts the discovery moment that ends fan relationships. A fan who already knew about the team and the tooling is not surprised by anything he reads on Reddit. A fan who was told every message was hand-typed and then discovers otherwise leaves immediately, often with a refund dispute attached.
What hurts revenue, consistently and significantly, is deception that gets found out. The chat operator who quits and posts screenshots. The agency that gets a competitor’s expose. The creator who says something on stream that contradicts what her account claimed. Every one of those is a slow-motion fan exodus.
Revenue is downstream of trust, not downstream of mystique. Fans pay for the perceived authenticity of the relationship, not for the belief that no tools were used to produce it. The creators who have run this experiment in public for the last three years agree on this point. Disclosure does not cost money. Discovery of hidden practices does.
A short close
You do not need to over-explain. You do not need to apologize. You do not need to list every vendor or describe every workflow. You need one or two clean sentences in your bio or your welcome message that tells fans what is true at the level they care about, and offers them a path to verify you personally if they want one.
Most fans will not even ask to verify. The offer alone is the trust signal. The disclosure alone is what keeps you on the right side of platform policy, the right side of your fans, and out of the Reddit threads that end careers.
If you work with an agency, ask them what their disclosure pattern is before you sign. If they cannot tell you in one sentence, that itself is information. If they tell you they claim every message is hand-typed, that is also information, and the information is that they are betting their creators on a deception that will be discovered eventually. You do not want to be one of the creators on the wrong end of that bet.
Pick the level you can live with. Write the line. Pin it. Move on. The disclosure question is one of the smallest hurdles in this work, and one of the most over-thought.