Taylor Swift not too long ago filed a sequence of trademark functions designed to guard the star from AI-enabled impersonations. Swift already holds a big selection of emblems, however these newest filings, at the very least one mental property agency suggests, serve a brand new goal: defending the timbre and character of her voice itself by means of what is named a “sound mark.”
In two current filings, posted April 24 by Swift’s firm, the movie star utilized to trademark two recordings. In a single, she says, “Hey, it’s Taylor,” and within the different, “Hey, it’s Taylor Swift.” The recordings themselves usually are not notably novel, however that’s seemingly irrelevant.
“The idea of defending sound as a trademark is just not new, although it stays comparatively uncommon,” wrote Josh Gerben, the Gerben IP legal professional that noticed the emblems on the regulation agency’s web site. “Traditionally, singers relied on copyright regulation to guard their recorded music. However AI applied sciences now permit customers to generate totally new content material that mimics an artist’s voice with out copying an present recording, creating a spot that emblems could assist fill.”
Gerben added that, in principle, if an AI-generated imitation of Swift’s voice grew to become the topic of litigation, she might argue that makes use of resembling her registered vocal emblems infringe on her mental property rights.
Gerben surmises that the aim is to guard the sound of Taylor Swift’s voice very like NBC protects its signature chimes. The technique, which Matthew McConaughey has also pursued, displays a novel strategy for the AI age, although it stays untested in court docket.
Celebrities are amongst these most susceptible to AI-enabled impersonations and broader unauthorized makes use of of their likenesses. Whereas prime artists and actors already face a permanent, whack-a-mole-style battle towards fakes, the most recent era of AI fashions has made producing these imitations unnervingly straightforward and scalable.
For comparable causes, celebrities, notably girls, are ceaselessly targeted by deepfake operations that use their faces and our bodies in nonconsensual pornographic imagery. Swift herself has been subjected to such campaigns, together with in early 2024, when illicit AI-generated photographs of her unfold extensively on platforms like 4chan.
In response, and for higher or for worse, celebrities are racing to put in guardrails of the AI age—or at the very least, making an attempt to determine learn how to construct them.
Swift’s try to guard herself from AI by way of sound marks is simply the most recent instance. In 2024, OpenAI paused the rollout of a ChatGPT voice that carefully resembled Scarlett Johansson’s—and, in an particularly recursive twist, her efficiency because the chatbot in Her—after Johansson publicly criticized the corporate for allegedly imitating her voice. (OpenAI has mentioned it used a distinct actor for the function.)
In one other instance, the household of Martin Luther King Jr. pressured OpenAI to take away likenesses of the civil rights chief from its video era platform, Sora, earlier than it was shut down.
And, little doubt underneath stress from expertise companies, YouTube not too long ago mentioned that it could broaden its deepfake detection service to Hollywood, and celebrities will now have the choice to request that sure movies that includes AI generations of them be.
“With assist from main expertise companies and administration firms, together with CAA, UTA, WME, and Untitled Administration, we’ve labored to refine how likeness detection can greatest serve expertise,” the platform mentioned in a statement. “We’re excited that celebrities and entertainers are actually eligible to entry this instrument, no matter whether or not they have a YouTube channel.”
In a market the place look and likeness are all the things, AI presents, at minimal, a brand new annoyance for artists in search of management, together with monetary management, over how their face and voice are used. That pressure will seemingly proceed to frustrate celebrities. Final 12 months, greater than 400 Hollywood leaders wrote to OpenAI and Google opposing using copyrighted work to coach fashions with out permission.
It’s notable that celebrities are pushing for protections towards a few of AI’s most noxious abuses. What stays unclear is whether or not these protections will prolong to the remainder of us, who additionally face the rising threat of digital impersonation, or just permit the Hollywood elite to decide out of a brand new web more and more full of countless uncanny mimicry.
