Taylor Swift trademarks her voice — and exposes the UK’s AI mimicry gap

TL;DR: TAS Rights Management has filed US trademark applications covering two spoken phrases — “Hey, it’s Taylor Swift” and “Hey, it’s Taylor” — alongside a stage photo of the singer. The strategy reaches for trademark law to fill gaps in copyright that AI-generated soundalikes exploit. IP scholars are sceptical the audio clips meet trademark-use thresholds. For UK artists, the story exposes a sharper problem: Britain has no equivalent statutory voice-rights regime.

The applications, which include audio specimens drawn from Amazon Music Unlimited promotion clips, target a specific weakness. Copyright protects the song, not the voice; AI models that mimic timbre and delivery without copying lyrics or melody slip through. Trademark — particularly soundmarks — could capture imitations that are “confusingly similar” rather than identical, IP attorney Josh Gerben told The Verge.

Northeastern University law professor Alexandra Roberts told The Verge she doubts the audio clip meets the use-as-a-mark requirement. Soundmarks have historically been narrow and isolated — the NBC chimes, the MGM lion roar — rather than phrases embedded in longer messages. UCLA’s Xiyin Tang suggested the registrations may function more as a deterrent: a federal certificate flashed at unsophisticated infringers, not a courtroom-ready weapon. Matthew McConaughey trademarked “alright, alright, alright” earlier this year on a similar logic.

Looking forward — and the UK angle

For UK-based artists looking at the same problem, the toolkit is thinner. The US patchwork includes state right-of-publicity laws, with Tennessee the first to address AI voice clones specifically (the ELVIS Act, 2024). The UK has no equivalent statutory voice right; passing-off and image-rights case law cover narrow circumstances and are slow to litigate. UK artists facing AI-generated soundalikes today rely on copyright-adjacent tools — producer-tag takedowns, performance-rights claims — rather than direct voice protection.

For UK SMEs in creative industries, particularly music, audiobook and voice-over, the read-across is that the legal scaffolding around AI mimicry is being built abroad first. The UK government’s stalled AI-and-copyright consultation was about training data, not voice cloning of finished output. Until that gap is addressed, contractual and platform-level controls — YouTube’s deepfake detection, label-side enforcement — are the practical shield.

Share this article