Targeted Chain of Attacks Aimed at Italian Public Figures
Unlike other cases covered in the article, the Italian attacks do not involve a one-time attempt to defraud a single firm. Rather, this chain of attempts is aimed at a certain target category: people around prominent Italian businessmen [1].
The method is simple and consistent: the fraudsters cloned the voice of a prominent person from publicly available interviews and other video material, and then proceeded to call some of his acquaintances and business partners. Typical script: “Something went wrong; please help”.
The Italian authorities have investigated cases of voice deepfake impersonation of Giorgio Armani and a member of the Beretta family, famous for the gun-making business [2]. Specific financial damage is undisclosed.
Why Prominent Personalities Are Especially Vulnerable Targets
For voice deepfaking any public person, the rule applies: the more public material available, the better the model can be trained. Giorgio Armani, a 90 year-old fashion entrepreneur of world renown, left numerous interviews and documentaries for public access.
Fraudsters don’t need a lab and any special knowledge. All that’s needed is 30 seconds of clear audio and publicly available software for deepfake generation.
The network of personal connections surrounding a prominent personality – friends, business partners and family – is targeted precisely because of the trust between them. Deepfakes use this trust.
A Global Trend Manifesting in Italy
Italian case is a manifestation of a global trend: personalized deepfake attacks targeting the personal networks of prominent individuals.
In the United States, the FBI warns of increasing vishing attacks with the use of cloned voice aimed at CEOs and financiers. According to the Chinese authorities, there were 20+ cases of impersonation of entrepreneurs through their voice clone [3].
Corporate level protection is simply not sufficient anymore. Deepfakes target the personal level: family, close friends, personal network of business connections. These areas remain completely unprotected.
“This was a voice of Giorgio Armani. Cloned and manipulated by fraudsters. Hard to notice. That’s why you should be careful”.
— Italian security expert, Corriere della Sera, 2024
How SYNHAWK Detects This Type of Attack
SYNHAWK PROTECTION: For such situations HAWK 7 and AURA ID modules of SYNHAWK can be used. HAWK 7 will detect an ongoing attack in the process of the call, monitoring audio communications. AURA ID will monitor public sources, raise an alert when someone tries to use a person’s digital image for manipulative purposes.