Cybercriminals do not need large resources: with a few seconds of audio, a well -built prompt or email templates fed by AI can perpetrate massive scams with unpublished precision and speed. According to Zerod’s specialists, the rhythm and realism of these campaigns force to rethink cyberdefense.

“With automated messaging systems, they can serve dozens of victims at the same time and generate communications that mimic the corporate tone of a real hotel. That cloning is used in calls to order urgent transfers or convince an employee to share sensitive information. The same technology that criminals use can be used to defend, but with the advantage of the speed and exhaustivity that only the AI ​​can offer,” Victor explains Ronco, CEO of Zerod.

MAIN VECTORS: Hospitality, politics and voice impersonation

Scams to the tourism sector

Tourism is among the most attacked sectors. Criminals are passed through travel agents or hotel receptions using automatic emails and chats that reproduce corporate styles. The objective: credentials and payment data. With the credentials committed, fraudulent purchases, false reserves or sales of data in clandestine markets are made.

Deepfakes and misinformation

Audio and video deepfakes have already been incorporated into misinformation campaigns. This type of audiovisual clips can erode the reputation of public figures and sow distrust in electoral processes. And, the ease of producing plausible material increases the risk of large -scale manipulation.

Voice cloning and prompt injection

A voice fragment is enough to create a convincing clone capable of ordering transfers (vishing) or raising human controls. On the other hand, the techniques of Prompt Injection (injecting malicious instructions into assistants or workflows of AI) open gaps in applications that were previously considered safe.

The AI ​​not only helps the attacker: he can also defend

Zerod and other ecosystem companies propose to use AI to get ahead of cyber attacks. Autonomous Penteting Tools allow not only to identify vulnerabilities, but exploit them in controlled environments to demonstrate the real impact. According to Ronco, that approach “shows exactly how a real attacker could compromise a system” and helps prioritize corrections.

Urgent measures to reduce risk

Strengthen training

Continuous staff training is essential: employees must recognize automated attempts and additional verification protocols.

Multiply the verification layers

Robust multifactor authentication, validations for independent channels and confirmation policies for economic transfers reduce the success of impersonation -based attacks.

Periodic drills and audits

Performing exercises that include Deepfakes and Vishing helps measure resilience and adjust procedures before a real incident occurs.

What do numbers say

Official statistics show an increase in digital incidents and online fraud, with thousands of cases managed annually by national response centers. Well, digitalization without robust controls facilitates the expansion of crimes based on AI.