Taylor Swift's company, TAS Rights Management, has filed sound marks for the spoken words “Hey, it's Taylor Swift” in the UK, the EU and Australia, extending US applications filed on April 24, 2026. None is a registration. As of October 15, 2026, both US voice marks are still waiting for an examiner, and only Australia has accepted its filing.
The overseas paperwork went in on October 6 and 7, as Music Business Worldwide first reported, and the timing is deliberate. The UK and Australian applications claim the April 24 US date as their priority date, and under the Paris Convention an applicant has six months from a first filing to do that abroad.
| Registry | What was filed | Status |
|---|---|---|
| USPTO | Two sound marks, “Hey, it's Taylor” and “Hey, it's Taylor Swift,” plus an image mark, filed April 24 | Sound marks awaiting examination |
| UKIPO | A sound mark (October 6) and an image mark (October 7) | Examination |
| EUIPO | A sound mark (October 6) and a figurative mark (October 7) | Under examination |
| IP Australia | A sound mark for “Hey, it's Taylor Swift” | Accepted October 12, in opposition period |
What exactly did Swift file?
Not her voice in general. US records (serial numbers 99784979 and 99784980) are sensory marks with no drawing, and each one says the mark “consists of the spoken words” of one short phrase. Both sit in Class 41, covering entertainment information and websites with non-downloadable audio and video about a musical artist. A third US filing covers an Eras Tour photograph of Swift with a pink guitar. The US papers include her signed consent to register her “likeness (including but not limited to signature, voice and image).”
Can a voice be a trademark?
A sound can. Netflix's “tu-dum” and NBC's chimes are the textbook examples, as attorney Josh Gerben noted when he flagged the US filings. His case for Swift rests on the confusion standard: “Trademark law doesn't just stop identical uses (like copyright law): it stops anything that is confusingly similar to the registered trademark.” In the same post he conceded that “attempting to register a celebrity's spoken voice is a new use of trademark registration that has not been tested in court before.”
The human voice does not fit as neatly within the trade mark framework as more conventional sound marks.Dr Basak Bak, University of Reading School of Law
A trademark identifies who stands behind a service; it does not own a timbre. A cloned Swift vocal that never says the registered words, used outside the listed services, is a far harder trademark case. “If you're palming something off as her, that's not OK,” David Herlihy, who teaches music and law at Northeastern, told the university's news service. “But if I try to sing like her, shouldn't that be OK?”
A filing is a claim. A registration is a right. A ruling is the only thing that says what the right is worth.
Others are betting anyway: MBW reports that Lionel Richie filed four US applications in June, each covering audio of him saying a phrase from one of his songs, and that the Backstreet Boys filed one for “Hi, we're the Backstreet Boys.”
What can an artist without Swift's legal team use?
For most artists the stronger tool is the right of publicity, which protects identity directly. Tennessee's ELVIS Act, signed in March 2024 and in force from July 1 that year, added voice to the state's protections. Law professor Jennifer Rothman warned: “It also creates some ambiguities and we may not know for years to come its boundaries.” Its first big test is human, not AI: in November 2025 the Johnny Cash estate sued Coca-Cola over a sound-alike singer in an ad, leaning on Bette Midler's 1988 win against Ford.
The federal fix is stuck. The NO FAKES Act of 2026 (S.4591), introduced by Sen. Chris Coons on May 20, would create a federal right in a person's voice and likeness. The Senate Judiciary Committee advanced it on June 18. On September 30, Sen. Ted Cruz blocked Sen. Marsha Blackburn's bid to pass it by unanimous consent, telling her, “at least as of now, I object.” The Daily Journal reported that the objection leaves the bill stalled for now; the identical House bill sits in committee.
Until a federal law passes, an independent artist has a state law that depends on where they live, the platform takedown route we covered in Sony's 260,000 deepfake removals, listener-side filters like Deezer's AI toggle, and the contract fight over who gets to say yes to a voice model. The next real signal is an examiner's: the first USPTO ruling on whether five spoken words can point to Taylor Swift.