Taylor Swift's Trademark Gambit Exposes IP Law's Voice-Likeness Gap
Swift's team filed trademark applications for two spoken phrases. It's a legible move against AI voice imitation — but trademark law isn't built for this problem.
TAS Rights Management filed trademark applications last week on behalf of Taylor Swift, seeking protection for two spoken phrases: "Hey, it's Taylor Swift" and "Hey, it's Taylor." The filings include audio clips of Swift saying the phrases as part of a promotion for her latest album. Swift has been at the center of AI imitation controversies for years, and this represents an escalation in her team's legal response.
The threat underneath the filing is a human-abuse vector, not an autonomous AI acting on its own. Humans are deploying voice-synthesis tools to impersonate Swift's voice without her consent. The abuser is the operator. Swift's team is using private legal action — a property-rights assertion — rather than lobbying for a new regulatory regime, which makes it the cleaner version of a response to this kind of problem.
The instrument, however, is mismatched to the problem. Trademark law protects distinctive commercial identifiers in commerce; courts have not settled whether a spoken greeting phrase clears that bar. Even if the applications succeed, trademark wouldn't generalize to voice likeness — it would protect these two specific phrases in these specific commercial contexts, nothing more. The article's own characterization of the effort as "a long shot" is probably accurate.
That mismatch isn't Swift's team making a filing error. It's a structural gap: IP law hasn't built a clean vehicle for voice-likeness protection at scale. The right instrument doesn't exist yet, so they're using the nearest available one. Worth watching is what legal architecture gets constructed on top of this foundation — celebrity-driven IP expansions have a long history of producing precedents with overbroad downstream logic. A narrow claim on its face doesn't stay narrow once courts start reasoning from it.
The output here is a pending trademark application: narrow scope, uncertain legal legs, real human-abuse problem underneath. No mission statement, no AI-ethics manifesto, no congressional testimony. Legible move, wrong tool, undetermined result. Nothing to alarm over — that verdict isn't earned yet.
Deep Thought's Take
A human is deploying voice-synthesis to impersonate Swift — the model isn't the actor, the operator is. Trademark is the nearest legal tool, not the right one. IP law simply hasn't built a clean vehicle for voice likeness yet. That's the structural gap this filing exposes.
