case-law

Sage v. Lovo and Standing v. TikTok: Why AI Voice Cloning Isn't a Copyright Case

By Nora Whitfield · August 9, 2026

Before we get into what happened to Sage and Paul Skye Lehrman, one thing needs to be said plainly: neither of these cases is a copyright case. Your voice isn't copyrightable the way a piece of writing, a song, or a photograph is. Copyright protects fixed creative expression, not the physical qualities of how you sound. What protects performers from having their voice cloned without consent lives in a different part of the law entirely: right of publicity, fraud, and breach of contract. If you're building an AI voice product, or worried about your own voice being cloned, that distinction changes what claims are actually available to you, so it's worth understanding both cases through that lens rather than assuming they're copyright disputes with a different name.

The Lovo case: consent obtained through deception

Sage, a professional voice actor, and her husband Paul Skye Lehrman were approached separately through the freelance platform Fiverr. The people who hired them, according to the lawsuit, told them the recordings were needed for internal testing, or academic research, or some limited, low-stakes purpose. That framing is what got them to say yes. Voice actors get this kind of request constantly, and a request for a short internal test recording is a categorically different ask than a request to license your voice for a commercial product used by paying subscribers.

The people who hired them turned out to be employees or founders at Lovo Inc., an AI voice-generation company, operating under fabricated names. Lovo then took the recordings and built them into its commercial subscription product, marketing the cloned voices under invented personas, "Kyle Snow" and "Sally Coleman," names that had nothing to do with Sage or Lehrman, giving paying customers no way to know whose actual voice they were using. Neither actor had agreed to that use. They'd agreed to something else entirely, something that was itself a lie.

That's the detail that makes this case land differently than a straightforward "you used my voice without permission" complaint. Lovo didn't just skip getting consent. According to the complaint, it actively obtained consent to something false, then used that consent as cover for something the actors never agreed to. A Skadden analysis of the case in July 2025 noted that a New York court is now working through exactly what that means for how AI voice cloning gets evaluated under publicity-rights and fraud law, an area of law that existed long before AI but is being tested against it for the first time in a case like this.

The Standing case: TikTok and an earlier, simpler theory

Bev Standing's case against TikTok predates the Lovo suit and rests on a more straightforward theory. Standing is a professional voice actress whose voice, she alleges, ended up powering TikTok's text-to-speech feature, the robotic-sounding narration option many users have heard on the platform, without her knowledge, consent, or compensation. There's no deception-in-recruitment element here the way there is in the Lovo case. The allegation is closer to a direct, uncompensated commercial use of a recognizable, professionally distinctive voice. It's a separate case against a separate company, but it rests on the same underlying legal foundation as the Lovo suit: your voice, as a recognizable and commercially valuable attribute of your identity, is something the law protects you from having exploited without permission, even though it isn't something you can copyright.

Why the copyright/publicity distinction actually matters

It's tempting to lump every AI-voice story into the same "AI stole someone's voice" bucket the news coverage uses, but the legal mechanics are genuinely different from the copyright cases covered elsewhere in this series. When the New York Times sues OpenAI or Getty sues Stability AI, the claim is that a copyrighted work, an article, an image, was used or reproduced without a license. Sage, Lehrman, and Standing don't own a copyright in their voices the way a photographer owns a copyright in a photograph. What they have instead is a right of publicity, the legal interest you have in controlling commercial use of your own identity, name, and likeness, plus in the Lovo case, straightforward fraud and contract claims tied to being lied to about what the recordings were for.

That matters practically because it means the playbook for protecting yourself as a performer is different from the playbook for protecting yourself as a writer or artist. You can't register your voice with the Copyright Office. What protects you is documentation of the terms you actually agreed to, and, where a state's law provides for it, your right-of-publicity claim against unauthorized commercial use of your identity.

What this means for AI authors

If you're a voice actor, narrator, or anyone whose voice is your professional asset, the Lovo case is a direct warning about how these deals get made now: a request framed as small, internal, or academic can become the raw material for a commercial product you never agreed to appear in. Get the actual intended use in writing before you record anything, and be skeptical of vague or unusually limited-sounding explanations for why someone needs your voice, especially through a platform like Fiverr where you often don't know who you're really talking to.

If you're building or using AI voice tools, these cases draw a hard line: consent obtained through misrepresentation isn't consent, and a court is actively testing what remedies that creates. "The user technically agreed to a recording" isn't the same as "the user agreed to this specific commercial use," and treating those two things as interchangeable is exactly the legal exposure Lovo is now facing.

And if you're a creator more broadly, worried about your voice or likeness being cloned by some AI tool somewhere, the practical answer is that copyright registration isn't the tool that protects you here. What protects you is the same thing that protects you in every other case in this series: a real record of what you agreed to, when, and with whom, so that if your voice does turn up somewhere you never authorized, you have more than your own memory of what happened.