How Creators Sell Their AI Voice Clones as Products

Lucas: So, a few months ago I noticed something in my TikTok feed: this creator I follow — she does skincare reviews, never shows her face — suddenly started posting these voiceover-heavy videos with a tagline that said 'trained on my voice.' And I clicked into her bio, and she had a link to something called 'Buy My Voice.' Luna: Wait — like, people could literally purchase her voice? Lucas: Exactly. She had partnered with a startup called VocalSynthetics. They take a sample of your speech — about thirty minutes of reading script — and they train a model that replicates your voice with about ninety-five percent accuracy. Then creators license that model to brands, to small businesses, even to other creators who want a voiceover but don't want to record it themselves. Luna: Hold on — so a brand could use her voice to narrate an ad, and she never has to step into a studio? Lucas: Never. And she never even has to approve the script line by line — the brand just types text and the AI reads it in her voice. She told me she made about forty thousand dollars in six months from that single deal. That's on top of her regular sponsorship income. Luna: That's wild. But isn't there a huge risk of misuse? I mean, what if someone generates something offensive or political using her voice? Lucas: That's the big question. The contracts typically have a content policy — you can't generate hate speech, you can't impersonate someone else, there's usually a review layer. But enforcement is spotty. VocalSynthetics uses a moderation API that flags certain keywords, but it's not perfect. Luna: And what about ownership? If the model is trained on her voice, does she own the model, or does the company? Lucas: In the standard deal, the creator retains ownership of their voice data — the raw recordings — but the trained model is jointly owned or sometimes fully owned by the platform. It varies. Some creators are negotiating better terms, especially the larger ones. A creator with a million followers can demand a higher royalty split. Luna: So what does the typical royalty look like? Is it a per-use fee, or a subscription? Lucas: Both models exist. Some platforms charge brands a monthly subscription — say, fifty dollars a month for up to a hundred thousand words — and the creator gets a percentage, usually between thirty and fifty percent. Others charge per generation, like five cents per generated audio clip, and the creator gets half. The skin-care creator I mentioned does a hybrid: she gets an upfront fee of two thousand dollars for the training, then thirty percent of every subscription sale. Luna: That's a pretty good revenue stream for something you only have to create once. But it also feels like it could commoditize voice acting. If every creator can sell their voice, won't the market get saturated? Lucas: It could, but there's a differentiation layer. Listeners build trust with a specific voice over time. If a skincare brand uses a voice that sounds like the creator they already follow, that ad has higher recall. So creators with an existing audience have an edge. It's not just about the voice quality — it's about the relationship. Luna: That makes sense. But what about the smaller creators who don't have a huge following? Can they still get into this? Lucas: Absolutely. There are micro-licensing marketplaces now. Fiverr actually launched a category last year called 'AI Voice Models' where you can list your voice for as little as ten dollars per commercial use. And some creators are doing it on Patreon — they offer a custom voice pack as a tier reward. For example, a gaming creator might let patrons generate short clips in their voice for a monthly fee. Luna: So it's an extension of the creator-audience relationship. Instead of just a shoutout, you get a personalized audio message. Lucas: Exactly. And the technology keeps improving. A year ago, you needed at least an hour of studio-quality audio to get a decent model. Now some platforms claim they can do it with just two minutes of casual speech from a TikTok video. That opens the door for way more creators. Luna: But doesn't that also make it easier to clone someone without their consent? I mean, two minutes of audio from any public video... Lucas: That's the dark side. There are already cases of deepfake voice scams using cloned audio of public figures. The FTC has been looking into it. For creators, the risk is that someone could scrape their content, train a model, and use it maliciously. Most platforms require explicit consent — you have to upload a video of yourself reading a consent script — but enforcement is hard. Luna: So what's the legal landscape? Can a creator sue if their voice is used without permission? Lucas: It's murky. In the US, your voice isn't protected by copyright — it's not a fixed work. But there are right-of-publicity laws in some states that give you control over commercial use of your likeness, which includes your voice. Tennessee actually passed the ELVIS Act last year, which explicitly protects voice against AI replication. But it's state by state. Luna: So if you're a creator, the smart move is to get in early, negotiate clear contracts, and probably watermark your audio somehow? Lucas: Watermarking is still experimental — there are inaudible digital signatures that can be embedded in audio, but they're easy to strip. The real protection is platform policy and legal recourse. The platforms have a responsibility to verify consent and take down unauthorized models. But it's an arms race. Luna: It feels like we're at the very beginning of this. I wonder how many creators are even aware that this is an option. Lucas: Not many. I polled my audience last week — about two thousand people — and only twelve percent knew you could license your voice as a product. That's a huge gap. The early movers are cleaning up. Luna: And you mentioned it's not just for big creators. What's the smallest example you've seen work? Lucas: There's a creator on TikTok with about fifteen thousand followers — she does ASMR baking videos. She never speaks in her videos, but she has a very distinctive soft voice that she uses for voiceovers on other platforms. She listed her voice on a site called VoiceSwap for five dollars per generated clip. In three months, she made about eight hundred dollars. That's real pocket money for someone who wasn't monetizing at all before. Luna: Eight hundred dollars for something that took her thirty minutes to record. That's a pretty good return on time. Lucas: And it scales. If she grows her audience, she can raise her price. Some creators are already charging fifty dollars per commercial use. The ceiling is high. Luna: So if you're a creator listening right now, the takeaway is: record a few minutes of clean audio, find a platform that offers voice cloning with consent verification, and put a price on it. Lucas: Exactly. And think about your niche. If you're a finance creator with a trusted, authoritative voice, that's valuable for explainer videos. If you're a comedian with a unique cadence, brands might want that for ads. Your voice is an asset. Luna: You know, hearing how creators are turning their voice into a product — it really makes you think about what other overlooked assets creators have. And honestly, if these marketing conversations have sparked something you've actually used, that's the kind of value we love hearing about. Lucas: Yeah, a lot of work goes into researching these angles and finding the actual numbers. If today's episode was worth a coffee to you, there's a link where you can toss one our way — it's buy me a coffee dot com slash fexingo. Luna: And it really does help keep the show ad-free and focused on the strategies that actually work. Lucas: Right. So back to voice clones — one thing I want to touch on is the ethical side for brands. If you're a brand using a creator's AI voice, you need to be transparent. The FTC has guidelines that say you should disclose if a voice is synthetic, especially if it mimics a real person. Failing to do that could land you in hot water. Luna: Has there been any enforcement yet? Lucas: A few warning letters. No major fines yet, but it's coming. The EU's AI Act has specific provisions for deepfake disclosure. So brands that jump in without clear labeling are taking a risk. Luna: So the smart play is to treat a voice clone like any other endorsement — disclose the relationship, disclose the synthetic nature, and get clear consent. Lucas: Exactly. And creators should be asking brands to include that disclosure in the contract. It protects everyone. Luna: What do you think the next evolution of this will look like? Maybe full digital avatars with cloned voices? Lucas: That's already happening. We're seeing creators license their entire persona — face, voice, mannerisms — as a virtual influencer. But that's a bigger episode. For now, voice alone is a low-friction entry point. It's one of the easiest digital products a creator can create. Luna: I think we'll see a lot more creators doing this in the next six months. The tools are getting cheaper and the audience is getting used to hearing synthetic voices. Lucas: And that's the key takeaway: if you have a distinctive voice and an audience that trusts it, you can turn that into a revenue stream with almost zero ongoing effort. That's the kind of passive income the creator economy promises but rarely delivers. Luna: Well, I'm sold. I might record my own sample this weekend. Lucas: Do it. And let us know how it goes.