Italian Prime Minister Giorgia Meloni has filed an application with the European Union Intellectual Property Office (EUIPO) to register a sound trademark for a short recording of her voice. According to Italian media, the goal is to protect against voice cloning by artificial intelligence and fake recordings. The idea sounds appealing, but trademark law was created to protect trade marks, not personal identities. Therefore, it's worth analyzing what such registration can achieve and where its effectiveness ends.
What exactly was reported?
According to information from the EUIPO register, cited by Italian media, the application was filed on October 5, 2026. It has the number 019431219, and the applicant is Giorgia Meloni, an individual. The application concerns an audio file lasting just under five seconds, in which the Prime Minister twice utters the phrase "Io sono Giorgia Meloni" ("I am Giorgia Meloni"). The application covers three Nice Classification classes:
- class 9: including downloadable and recorded multimedia content;
- class 41: including Cultural events; Conducting conventions;
- class 45: including organizing political meetings.
The mark has not yet been registered. The application is currently being examined by the EUIPO. If it passes, it will be published, triggering a three-month period for holders of earlier rights to file oppositions. Similar applications were previously filed in Italy by voice actor Luca Ward and singer Giusy Ferreri.
Can a voice be a trademark?
Since 1 October 2017, European Union trademark regulations, currently contained in Regulation 2017/1001, no longer require a trademark to be represented graphically. Any sign, including a sound, can be a trademark if it meets two conditions:
- Distinctiveness: The sign must be capable of distinguishing the goods or services of one trader from those of other traders.
- Clear and precise representation in the register: A sound mark is filed in the form of an audio file that anyone can listen to in the EUIPO database.
Sound trademarks are nothing new. Advertising melodies and sound signals have been registered for years. A relatively new phenomenon, however, is registering the speech of a specific person as a trademark. However, such a trademark does not protect the "voice" itself. Only the specific recording listed in the register is protected.
What does a sound sign give?
Registration grants the owner exclusive rights throughout the European Union for 10 years from the date of filing, with the possibility of renewal for subsequent periods. The rightholder may prohibit others from using a sign identical to the mark for identical goods and services in trade. They may also prohibit the use of identical or similar signs for identical or similar goods and services if there is a risk of confusion. Only renowned marks enjoy broader protection. The right under registration can only be asserted from its publication, and actions taken after the application is published are entitled to a specified amount of compensation.
Why won't a trademark stop deepfakes?
The limitations of this tool are significant, and they should be the main lesson from this case. Trademark registration law protects against the use of a sign in business. A fake recording distributed for political, disinformational, or humorous purposes is usually not considered use in business. A deepfake, in which a cloned voice utters completely different words, will most likely not be identical or similar to the registered mark. The EUIPO guidelines on sound marks containing words emphasize the verbal aspect of the comparison. Mere similarity of voice and intonation is generally not sufficient. If the mark is not genuinely used for the registered goods and services for an uninterrupted period of five years from registration, the rights to it may be deemed expired.
EU and Italian solutions
At the European Union level, the transparency obligations under Article 50 of the Artificial Intelligence Act (AI Act) have been in force since August 2, 2026. Anyone who professionally uses an AI system to create deepfakes, including audio recordings, must disclose that the content was artificially generated or manipulated. For works that are clearly artistic or satirical, this obligation is relaxed, and a violation can result in a fine of up to €15 million or 3% of the total annual global turnover. However, the AI Act does not prohibit the creation of deepfakes, and the labeling obligation does not cover purely private activities. This is supplemented by procedures for reporting illegal content, which online platforms are required to provide under the Digital Services Act (DSA).
In summary, registering a voice as a trademark is an interesting instrument, albeit one with limited protection options. It protects a specific recording from unauthorized use in business transactions, but it does not grant exclusive rights to the timbre of the voice and will not prevent disinformation. Effective protection against voice cloning requires a combination of several tools: protection of personal rights, personal data protection regulations, criminal law, obligations arising from the AI Act, and procedures for reporting content to platforms. The case of the Italian Prime Minister, however, may spark a discussion on how Polish lawmakers should regulate voice protection.
This article is for informational purposes only and does not constitute legal advice.
The law is current as of October 22, 2026.
