Giorgia Meloni’s most recognisable line began as a declaration of identity. Now Italy’s prime minister wants it to function as a legal barrier against artificial intelligence.

The phrase dates to a 2019 political rally, where Meloni said: “I am Giorgia, I am a woman, I am a mother, I am Italian, I am Christian.” It later became the title of her 2021 autobiography and the basis for a disco song, turning a campaign moment into a durable part of her public persona.¹

On Oct. 5, Meloni applied to the European Union Intellectual Property Office to trademark a four-second recording in which she says, “Io sono Giorgia Meloni” — “I am Giorgia Meloni.” The filing covers areas including downloadable multimedia content, cultural activities and political-event organisation, but remains under examination; no trademark protection is in force unless the EUIPO approves it.¹

Her case is driven by a sharper concern than branding. Generative AI can now produce increasingly convincing versions of a person’s voice and image, blurring the line between authentic material and manipulation. Meloni has already brought a pending civil case against a man accused of making and distributing inappropriate deepfake images using her likeness.¹

“I can defend myself, but many others cannot,” Meloni said in response to those images, warning that deepfakes can “deceive, manipulate, and harm anyone.”¹ Italian authorities have faced the problem beyond the prime minister: in 2025, scammers used an AI imitation of Defence Minister Guido Crosetto’s voice to target wealthy businesspeople.¹

Meloni is not alone in testing the limits of identity protection. Performers including Taylor Swift have sought trademark protections involving voice and likeness, while disputes and warnings involving Scarlett Johansson, Tom Hanks and AI-generated celebrity material have exposed the same vulnerability.² Her EU application turns that wider anxiety into a concrete experiment: whether a voice can be guarded before a fake uses it against its owner.