Audio deepfake: only 3 seconds of voice and €30

Meta description: Audio deepfakes can be created from just three seconds of real voice at a low cost (about €30). Frauds, voice cloning, and misuse of celebrities are on the rise.

The technology capable of reproducing human voices with fidelity from only three seconds of recording and at an indicated cost of around €30 has been expanding fraud, vocal cloning, and misuse of celebrities, according to an ANSA report citing security companies and sector experts. In addition to legitimate applications, such as accessibility in cinema, experts warn of increasing scams through social engineering and targeted attacks on financial institutions and public figures.

What the report states

• From three seconds of real audio, it is possible to create a convincing replica of a person’s voice; the average cost to generate the synthetic voice can be around 30 euros, according to the ANSA article.
• The technology has positive uses — for example, for voice recreation in cultural and accessibility projects — but it also facilitates fraud and manipulation.

  "Audio deepfake can be created with three seconds of voice and cost around €30", according to ANSA.

Cases and examples cited

• Legitimate use: recreation of actor Val Kilmer’s voice in a film, cited as an example of application for artistic and accessibility purposes when the technology is used to recover lines for actors with vocal limitations.
• Frauds: the report mentions an attack involving voice cloning used against a bank manager at Fideuram, in addition to manipulations involving celebrities, where fake voices were used to create deceptive content or pressure targets financially.
• Institutional reaction: Prime Minister Giorgia Meloni registered her voice with the European Union Intellectual Property Office (EUIPO) as a preventive measure to defend against AI-generated forgeries, according to ANSA.

Data and trends according to the sector

• The ANSA report, citing data from cybersecurity firm McAfee, shows significant growth of the phenomenon:
  • 900% increase in deepfakes (video, photo, and audio) in one year.
  • Fraud attempts grew by more than 3,000%; attacks rose 700% year over year in 2026.
  • Crimes linked to vocal cloning and audio deepfakes rose by 1,300%, according to sector figures cited.
• These metrics indicate not only greater technical capability but also commercial and criminal scalability of cheap voice synthesis.

Risks and implications

• Criminals use vocal cloning for social engineering scams: fake transfer requests, instructions to subordinates, or pressure on victims’ relatives.
• The availability of low-cost synthesis services lowers barriers to entry for attackers, increasing the number and sophistication of frauds.
• Beyond financial loss, there is reputational risk for public figures, companies, and institutions — fake voices can be used to manipulate opinion, extort, or influence decisions.
• Detection technologies are evolving, but specialists warn that multifactor authentication and additional verification protocols remain essential.

Relevance for Brazilians and descendants of Italians

• The threats of audio deepfakes are not bound by borders: Brazilians and communities of Italian descent can be targets of similar scams, especially where there are cross-border financial, bureaucratic, or personal relations.
• Financial companies, managers, and public figures in Brazil face risks similar to those cited in Italy; voice call or voice-recorded phishing scams have already been reported in several countries.
• Recommended preventive measures include:
  • adopting multifactor authentication for sensitive operations;
  • internal protocols requiring confirmations through separate channels (for example, written message or in-person verification) before authorizing transfers;
  • recording and documenting critical communications;
  • attention to typical deepfake cues (odd intonation, unusual pauses, contextual inconsistencies).
• For those following Italian citizenship processes or maintaining contacts in Italy, it is prudent to verify identities through official channels and be alert to financial requests by voice or audio. News on the topic and official Italy-related measures can be followed in Italy News and in sections on Life in Italy.

Conclusion

The combination of short voice samples and low costs to generate synthetic audio makes audio deepfakes a growing threat to cybersecurity, financial fraud, and protection of public image. While the technology offers legitimate benefits, experts and authorities recommend preventive measures and robust verification protocols to reduce risks.

Source: ANSA