Are AI influencers ethical? Where we draw the line

An AI influencer can be run ethically if 3 things hold: her face is original, she’s labelled as AI where people will actually see it, and nobody is ever told she’s a real woman. The worst cases reported so far, from stolen videos with AI faces pasted on to fans who thought they were chatting with a real person, break at least one of those. This is the view we teach from, and it sits under our guide to making money with an AI influencer.

On this page
  1. What makes an AI influencer unethical?
  2. Is it deceptive to run an AI influencer at all?
  3. Do fans know AI models aren’t real?
  4. Is it OK to hire a chatter for an AI character?
  5. Do AI influencers hurt real creators?
  6. Where do we draw the line?
  7. What can go wrong?

What makes an AI influencer unethical?

Deception and theft. An AI character becomes unethical when she’s built from someone else’s face or body, when she’s passed off as a real person, or when fans pay because they believe they’re talking to one. Each of these has already happened, been reported, and cost real people something. The table pairs each practice with a documented case.

Practice Who gets hurt Reported case
Pasting an AI face onto a real woman’s video The woman whose body, work and content were taken AI Instagram accounts swapped AI faces onto real models’ and creators’ videos and linked to paid pages. Fanvue temporarily removed one account after 404 Media asked, saying it breached its terms (404 Media, 9 Apr 2024)
Letting followers believe she’s real Fans, and anyone who acts on what they think they saw An AI character’s Wimbledon posts were taken as real; her captions implied she was there (The Telegraph via Yahoo, 19 Jul 2025; IBTimes UK, 11 Jul 2025)
Chatters writing as if they were the creator Fans who pay for what they think is a personal connection Chatters in the Philippines writing as OnlyFans creators, and agencies training AI bots on their chat logs (Rest of World, 20 Aug 2025)
Using a real person’s likeness The person whose face it is Illegal in many places, and platforms restrict it: see AI likeness laws
Paid posts without an ad label Followers who think it’s an honest opinion The FTC’s Endorsement Guides cover endorsers who “could be or appear to be an individual”

All sources checked 4 October 2026.

Is it deceptive to run an AI influencer at all?

Not if people know. A fictional character is no more a lie than a cartoon mascot: Brazilian retailer Magazine Luiza has used a virtual spokesperson, Lu, since 2003. Deception starts when the AI part is hidden. The platforms draw the same line: they allow AI characters and require a label people can see.

As of 4 October 2026:

  • Instagram asks profiles featuring an AI-generated person to use its AI-generated profile label, and says those that don’t may see limits to their reach (Instagram, 31 Aug 2026).
  • Fanvue requires AI content to be disclosed clearly and prominently, in the bio, the captions or a watermark (Fanvue help center).
  • TikTok requires a label on AI content showing realistic people or scenes. Our TikTok AI label guide has the details.
  • The EU AI Act requires deepfake content to be disclosed from 2 August 2026, covered in EU AI Act labels.

Do fans know AI models aren’t real?

Some don’t. In July 2025 an AI character called Mia Zelu posted photos that appeared to show her at Wimbledon, and many followers treated her as a real fan in the stands. The Telegraph reported her bio disclosed AI; IBTimes UK said the bio “now” does. Either way, her captions implied she was really there.

The same pattern shows up with deliberate realism. Euronews reported in 2023 that a well-known actor messaged Aitana López, an agency’s AI character, to ask her out, and quoted the agency: “He had no idea Aitana didn’t exist.” More on her in our Aitana López case study.

So put the label where it can’t be missed: Instagram’s profile label switched on, an AI line at the top of the Fanvue bio, and AI mentioned in captions where it fits. Don’t write captions that imply she was somewhere real. If someone still asks, give a straight answer.

Is it OK to hire a chatter for an AI character?

Yes, as long as the chatter never claims she’s a real woman. A fan of a labelled AI character already knows she isn’t human; a real person writing her replies is part of how the character works. The line is crossed when anyone tells a fan “this is really me”, or lets him believe it to close a sale.

That’s the difference from what Rest of World reported in August 2025: chatters writing as real OnlyFans creators, to fans who think they’re talking to the woman herself. Rest of World also reported that OnlyFans’ rules require a human to send the messages, and that chatters’ logs are being used to train AI chatbots.

Our rule: if a fan asks whether she’s real, tell him the truth. She’s an AI character, which her profile already says, and a real person reads and answers his messages. The exact wording we use is in Fanvue chatting tips, and when to bring in help is in hiring a chatter.

Do AI influencers hurt real creators?

Sometimes, and it’s fair to say so. The founders of the agency behind Aitana López told Euronews and Fortune in 2023 they built her to stop depending on influencers’ fees and egos. That is work a real person used to be paid for. Stealing a real creator’s video and pasting an AI face on it does far more direct harm.

Other criticisms are fair too. Shudu, a CGI model, was criticised because a white man had created a Black woman; her creator told Business Insider in 2021 that writer Ama Badu became her voice and an advisor on the character. Aitana’s agency, answering criticism of her images, told Euronews that “the world in general is sexualised”. It’s a fair description of the market; it doesn’t settle the criticism.

What you can control is your own account. The course’s prompts ask for a candid smartphone photo with natural skin texture, not a flawless render, and never describe her body. Characters are always clearly adult. Dance videos start from a clip you filmed yourself or have a licence for, never someone else’s.

Where do we draw the line?

These are the seven rules this site and our guides follow. They aren’t legal advice; they’re how we think characters should be run so that nobody is deceived and nobody’s face, body or work is taken. If a tactic needs one of them broken, we don’t recommend it. The law side is in our AI influencer rules.

  1. Original faces only. No real person’s face or photos, no face swaps, no scraping social media for a dataset.
  2. Clearly adult. Every character looks and is written as an adult.
  3. Labelled everywhere. Instagram’s label on, an AI note in the Fanvue bio, TikTok’s AI label on realistic posts.
  4. Never claim she’s human. Not in the bio, not in captions, not in chat.
  5. Chatters tell the truth when asked. A real person answers, and the fan can know that.
  6. Your own source material. Poses and dance clips come from her own images, licensed stock, or clips you filmed yourself.
  7. No fake proof. No invented income, testimonials or reviews. We don’t publish income claims at all.

If you want to start this way, the first step is an original face: the free video 1 tutorial shows how.

What can go wrong?

  • You switch the label off to get more reach. It’s deception, and Instagram may limit the reach of unlabelled AI profiles anyway. The label stays on.
  • You use someone’s reel for a dance video. That’s the 404 Media story. Film the move yourself or license the clip.
  • A chatter says “it’s really me”. You lose the fan when he finds out, and it undoes the AI disclosure Fanvue requires.
  • Her face looks like someone famous. Followers will say so. Change the face before you train her LoRA, not after.
  • You post her “at” a real event and let people believe she was there. That’s the Mia Zelu pattern. Keep the fiction clearly fiction.

Questions people ask

Is it ethical to make money from an AI influencer?
It can be. Selling content from a fictional character is no different in principle from a brand mascot or a cartoon, as long as buyers know she's AI. It stops being ethical when the face belongs to a real person, when the label is hidden, or when a fan is told she's a real woman.
Do I have to tell fans my AI influencer isn't real?
Yes. Instagram has an AI-generated profile label and may limit the reach of AI profiles without it, Fanvue requires clear AI disclosure in the bio, captions or a watermark, and TikTok requires labels on realistic AI content. If a fan asks directly, tell him the truth.
Is AI OFM ethical?
It depends on how it's run. An original, clearly adult character, labelled as AI, on a platform that allows AI creators, with chatters who never claim she's human, is honest. Stolen videos with an AI face swapped in, or fans led to believe they're talking to a real woman, are not.
Do AI influencers take work from real models?
Some do. The founders of the agency behind Aitana López said they made her to stop depending on human influencers. That's a real cost of the technology. What you control is not stealing real creators' content and not passing your character off as a person.

Read next