Meloni’s voice trademark highlights new frontier in AI-era brand protection

Italy’s prime minister has applied to register a four-second recording of her voice as an EU sound mark. The filing raises a broader question for brand owners: can trademark law provide an additional layer of protection as AI makes convincing voice impersonation increasingly easy?

Italian Prime Minister Giorgia Meloni’s application to register a recording of her voice as an EU sound mark has attracted attention because of its unusual subject matter. But the filing also arrives at an important moment for European intellectual property and AI regulation.

Meloni filed the application with the European Union Intellectual Property Office (EUIPO) on October 5. It covers a roughly four-second audio recording in which she says, in Italian, “Io sono Giorgia Meloni” (“I am Giorgia Meloni”). The application is currently under examination.

The application, identified by EUIPO as file no. 019431219, covers goods and services in Classes 9, 41, and 45. These include downloadable multimedia content, cultural activities, and the organization of political events.

Meloni’s office has linked the filing to concerns over the use of generative AI to clone voices and create misleading content. The phrase chosen for the recording is also closely associated with Meloni, who used “Io sono Giorgia” during a 2019 political rally and subsequently used it as the title of her autobiography.

But describing the application simply as “trademarking her voice” risks obscuring an important point about what trademark law actually protects.

A sound mark is not blanket ownership of a voice

EU trademark law permits signs consisting exclusively of a sound or combination of sounds to be registered. EUIPO’s current guidance states that a sound mark must be represented by an audio file reproducing the sound or by an accurate representation in musical notation. The Office’s examples expressly include a voice saying a word. That does not mean, however, that a registration would give Meloni a general property right over every recording or imitation of her voice.

The scope of an EU trademark is determined by the sign that is registered and the goods and services for which protection is obtained. EU trademark rights principally address use of protected signs in the course of trade, subject to the conditions established by the EU Trade Mark Regulation.

Accordingly, a registered sound mark could potentially provide an additional enforcement tool where an unauthorized commercial use falls within the scope of the registration. It would not automatically mean that every AI-generated imitation of Meloni’s voice constitutes trademark infringement.

Trademark law meets the AI Act

Meloni’s application also comes shortly after an important development in EU AI regulation. Article 50 of the EU AI Act began to apply on August 2, 2026. Among its transparency requirements, deployers of AI systems that generate or manipulate image, audio, or video content constituting a deepfake must disclose that the content has been artificially generated or manipulated, subject to the conditions and exceptions set out in the regulation.

The European Commission says the relevant obligation applies where AI-generated or manipulated content would falsely appear to a person to be authentic or truthful. Disclosure must be made to a natural person at the latest upon first exposure and must be clear and distinguishable.

The AI Act and trademark law therefore address different problems. The AI Act is concerned principally with transparency: making clear when certain content has been artificially generated or manipulated. Trademark law, by contrast, is concerned with distinctive signs, commercial use, and the protection of source-identifying functions.

The two regimes could consequently operate alongside one another. An AI-generated recording could trigger transparency obligations under the AI Act while, depending on the circumstances and manner of use, also raising questions under trademark law or other areas of law.

The enforcement question

That makes Meloni’s filing less a new form of ownership over human voices than an example of a rights holder seeking to add another layer to the protection of a recognizable identity. The practical challenge is that AI-generated content can be created and distributed rapidly, while trademark enforcement still depends on establishing that the relevant use falls within the scope of the applicable rights.

A successful registration would therefore not prevent voice-cloning technology from existing. Nor would it necessarily prevent every imitation, parody or other use of Meloni’s voice. Its potential significance lies instead in whether the registered sound can be used effectively against particular commercial uses that satisfy the requirements for trademark infringement.

An emerging brand-protection tool?

Meloni’s application is not evidence that trademark law has suddenly become a general-purpose right in a person’s voice. EUIPO has long permitted sound marks, including marks consisting of spoken words. What is changing is the technological environment in which those rights operate.

Generative AI can now reproduce recognizable voices with increasing ease. That creates a new reason for individuals and businesses whose voices have commercial value to consider whether a distinctive recording, phrase, or other sound could function as a trademark alongside whatever protection may be available through personality rights, copyright, unfair competition, or other legal regimes.

There is already evidence of interest in this approach. ANSA reported that Italian singer Giusy Ferreri had also pursued protection for her voice through an EUIPO sound mark application.

The more significant question is therefore not whether a voice can be a trademark. EUIPO’s rules already establish that a sound can qualify as an EU trademark. And how useful that right will prove when the threat is not simply unauthorized use of an existing recording, but the synthetic recreation of a recognizable voice.

For Meloni, the answer will depend on whether the application proceeds to registration and, ultimately, how the resulting right could be enforced against particular uses. For brand owners, however, the filing highlights a broader development: as generative AI makes voices, images, and other recognizable identity cues easier to reproduce, the boundaries between brand assets and personality assets are becoming increasingly important from an IP perspective.

Meloni’s four-second sound mark therefore offers an unusually timely example of how conventional trademark protection could sit alongside the EU’s emerging regulatory framework for synthetic media.

← Back to News