<\/p>\n
When Taylor Swift’s intellectual property firm filed trademark applications in April 2026 covering short audio clips of her voice and her visual likeness, the move was widely read as a celebrity legal manoeuvre. Legal scholars say it reveals something more consequential: a structural gap in artificial intelligence law that copyright protections alone are not equipped to close.<\/p>\n<\/div>\n
The assessment comes from Daryl Lim, Associate Dean for Research and Strategic Partnerships at Penn State, writing in Fortune, who argues that Swift’s filings signal a fundamental shift in how the law is being asked to respond to AI-generated identity fraud, one that reaches well beyond the debate over copied songs and scraped datasets.<\/p>\n<\/div>\n
Swift’s company, TAS Rights Management, filed the trademark applications in April 2026 following a sustained pattern of AI-generated misuse. Fabricated nude imagery of the singer has spread across the internet, whilst her voice and likeness have appeared in manufactured political messaging and counterfeit product endorsements she never sanctioned.<\/p>\n<\/div>\n
The filings do not concern her recordings, lyrics or albums. They are aimed at a different legal problem: whether AI-generated versions of her voice or image could deceive the public into believing she has endorsed a product, a political position or a cause she has never approved.<\/p>\n<\/div>\n
The majority of AI-related litigation to date has centred on copyright law, which shields creative works, including songs, books, photographs, recordings and journalism, from being copied, distributed, adapted or performed without authorisation.<\/p>\n<\/div>\n
The pattern is well established in the courts. The New York Times sued OpenAI and Microsoft in 2023, alleging that the companies trained their AI systems on its journalism, generating outputs that competed with or reproduced the outlet’s reporting. Authors, publishers, photograph agencies and music publishers have brought parallel claims against other AI companies.<\/p>\n<\/div>\n
But copyright does not automatically protect a person’s identity. It does not grant Swift a general right to control anything that sounds like her, looks like her or calls her to mind in the audience’s imagination.<\/p>\n<\/div>\n
The critical limitation, according to Lim’s analysis, is this: if an AI-generated voice imitates Swift without copying a specific recording, song or lyric, copyright law may not address the genuine harm, which is the false impression that she said, sang or approved something she never did.<\/p>\n<\/div>\n
Trademark law starts from an entirely different set of concerns. It protects names, images, sounds and other identifiers that help consumers establish who or what stands behind a product or service.<\/p>\n<\/div>\n
Familiar examples include brand names such as Coca-Cola, logos such as the Nike swoosh, slogans such as Subway’s “Eat Fresh” and distinctive sounds such as the MGM lion roar.<\/p>\n<\/div>\n
A trademark is not a blanket ownership claim over a word, phrase, voice or image. It is a legal tool for ensuring that consumers know who stands behind what they are buying, hearing or seeing.<\/p>\n<\/div>\n