When Taylor Swift began re-recording her first six albums, she turned a dispute over master recordings into a public lesson in intellectual property ownership. Her fans didn’t need to understand every aspect of music copyright or the contractual history behind her original recordings to know Swift was making a major statement. She gave fans a clear reason to choose her versions, and it became a turning point for artists globally.
Swift’s latest trademark applications suggest that her strategy is expanding beyond creative works to the recognizable features of her identity. In April 2026, TAS Rights Management filed applications covering recordings of Swift saying, “Hey, it’s Taylor Swift” and “Hey, it’s Taylor.” A third application involves a moving image of Swift. Variety reported on the filings shortly after they were submitted.
These applications do not give Swift exclusive rights to her voice or appearance in every context However, this filing shows that her team is planning ahead and strategizing her commercially identifiable assets and persona.
Taylor’s story is not merely a celebrity branding story. It raises a question that should concern any executive whose company derives value from a recognizable founder, distinctive customer experience or trusted public identity: what causes the market to recognize your business, and how much control do you have over those elements?
AI can reproduce recognition without authenticity
Businesses have historically organized their IP portfolios around identifiable assets like company names, product names or visual logos. They traditionally protect inventions and creative works and designs where appropriate, establishing ownership with agreements.
Those measures remain necessary, but AI exposes a category of value that many portfolios do not fully address. A person or brand can now imitate the voice of a founder, reconstruct the appearance of a spokesperson or recreate the familiar presentation of a brand without copying any single asset in its entirety. The result may still feel authentic because audiences recognize the person or experience.
Swift has experienced the consequences directly through AI-generated images that have falsely connected her to political endorsements. Fake videos have appeared to show her promoting fraudulent offers, and in 2024, sexually explicit AI-generated images spread so widely that X temporarily restricted searches for her name.
The scale of Swift’s public profile makes her an obvious target, but the same methods can be directed at businesses of any size. For example, a fabricated video could show a founder announcing an investment opportunity or endorsing a product, or a competitor could mimic the language, visual presentation and product experience of a young company closely enough to benefit from the trust that company has built.
In each case, the impersonator is borrowing recognition without providing authenticity. The harm is not limited to lost sales or a single confused customer: rather, the harm can undermine confidence in legitimate communications, complicate commercial partnerships and diminish the licensing value of the identity being copied.
Identifying a company’s recognition assets
When our firm begins working with a company, our goal is to understand how the business creates value before deciding what its IP strategy should include. The registration schedule tells us what the company has already identified and protected. It does not necessarily tell us what customers have come to recognize or what will matter most as the business grows.
I want to know which features competitors consistently imitate and whether a founder’s public profile has become a source of credibility for the company. I consider whether customers recognize a product by its configuration, packaging or method of presentation before they see the name attached to it.
I think of these features as recognition assets. They are the elements that signal the source of a product, experience or message, including situations in which the company’s formal name or logo is not present.
For Swift, recognition comes from an unusually broad collection of assets. Her voice, signature, lyrics, visual eras, tour imagery and fan-facing phrases reinforce one another. Her audience of mega-fans has learned to read those signals, which allows individual elements to carry commercial meaning of their own.
A company’s recognition assets may be less famous, but they can be just as important to the business. They might include a distinctive product feature, recurring content format, founder-led video series or terminology developed to describe a proprietary method. For example, for a software company, recognition may come from the way the product communicates with users. For a consumer brand, it may come from packaging or an unusual demonstration that customers now associate with the product.
There is no single legal tool for protecting all of these assets at once. Depending on the brand or company, the strategy may involve trademark, copyright, design protection, contractual rights or rights of publicity. Certain assets may not qualify for registration at all, which makes the company’s agreements and commercial practices even more important.
The legal analysis can only begin once the business knows what it is trying to protect. This is where many companies fall behind. Their portfolios reflect the products and brand architecture they had at launch, even though the source of commercial value has changed. The founder may have become a prominent public voice, a product feature may have acquired independent recognition, or the business may have developed proprietary data and content that now support new revenue streams.
If the IP strategy is not revisited as the business evolves, the company may discover that one of its most valuable assets is poorly protected, owned by the wrong party or subject to agreements that never anticipated its current use.
Control should support innovation
Matthew McConaughey’s recent trademark strategy provides a useful comparison. His company secured registrations involving recognizable audio and visual elements of his persona, including his delivery of “Alright, alright, alright.” His legal team has explained that the purpose is to establish clearer boundaries around ownership and consent before his likeness is misused by AI. The Wall Street Journal reported on the strategy and the approved registrations.
McConaughey has also authorized an AI-generated version of his voice for a Spanish-language edition of his newsletter through ElevenLabs. That authorized use helps clarify the commercial objective behind the registrations. He is not trying to prevent AI from interacting with his identity under any circumstances, instead he is establishing a stronger position from which to decide which uses he will approve.
Executives should view protection in the same way. The ability to stop misuse is important, but ownership also supports licensing and expansion. For example, I would never recommend a company or figure misses out on opportunities, but there should be protection in place to authorize use of familiar assets and protect ownership.
Trademark registration may contribute to that position when a voice, motion or other nontraditional element functions as an identifier of source. The USPTO recognizes that sounds, motion, colors and other nontraditional features may function as trademarks. Registration does not create ownership of a human characteristic for every conceivable purpose, and the use of these registrations against AI impersonation remains relatively untested.
Effective protection may therefore require several overlapping rights. The appropriate response to a fake endorsement could involve trademark or false endorsement claims, while an unauthorized digital replica might also implicate publicity rights, contractual restrictions or newer laws addressing synthetic identities. Platform procedures may be necessary when speed matters more than litigation.
Any adviser promising that one registration will solve AI impersonation is oversimplifying both the problem and the solution. The legal perimeter must reflect the asset, the way it is being used and the specific commercial harm the company needs to address.
Protecting the company’s future value
An IP review should examine the business the executive team expects to operate several years from now, rather than documenting only what the company owns today.
If a founder is becoming an important public representative of the business, the company should understand what it owns and what remains personal to the founder. If customers recognize a product by its shape or presentation, the company should assess whether those features can be protected before entering additional markets. If proprietary material is being provided to an AI vendor, the agreement should address whether that material may be retained or used to improve systems serving other customers.
These issues become harder to resolve after the asset has gained value. A missing assignment from an early contractor may appear minor until the company enters acquisition diligence. Informal permission to use a founder’s image may become a serious problem when the founder leaves. Broad AI terms accepted from a vendor may conflict with assurances the company later wants to give its customers or investors.
Before expanding an IP portfolio, executives should understand what customers recognize as uniquely connected to the company, whether the company has meaningful control over those elements and how emerging technology may change their use. The answer will not always be another filing. It may require revised agreements, internal AI rules, clearer licensing terms or a system that helps the public distinguish official content from imitations.
That last point may be the most important lesson from Taylor’s Version. Swift did more than pursue control over her recordings. She taught her audience how to identify the versions she wanted them to support. As synthetic media becomes more convincing, companies may also need to show customers how to distinguish an authorized message, product or experience from one that merely reproduces familiar signals.
AI has made it possible to manufacture recognition at a speed and scale that existing IP portfolios were not designed to address. Businesses should identify the features that carry their credibility before those features are copied, licensed or incorporated into systems without their consent. By the time the market has been trained to accept the imitation, the legal options may be narrower and the commercial damage much harder to reverse.
