Bulgarian companies are becoming victims of large-scale cyber scams, in which millions of euros disappear every week. This was reported by the press center of the Bulgarian National Radio, referring to an official statement by Commissioner Svetlin Lazarov, head of the "Cyber Intelligence" department at DGOB. According to official data, in our country, business correspondence of companies is regularly attacked, with the amounts of a single fraud reaching between one and five million euros.
The main weapon against domestic business remains the replacement of international bank accounts (IBAN) and the direct compromise of company emails. Commissioner Svetlin Lazarov points out that every week between three and four Bulgarian companies fall into the trap of digital criminals.
For ordinary consumers in the country, the biggest threat continues to be phishing. Currently, fake electronic messages about unpaid fines to the Traffic Police are being widely distributed, which aim to steal personal data and bank cards.
At the international level, law enforcement agencies have achieved serious success by dismantling the specialized information network "First VPN". According to Europol, the platform offered hackers complete anonymity and a secure environment for organizing ransom attacks, advertising itself as a "jurisdiction-free" service.
The joint operation, in which 18 countries participated, achieved the following results:
33 servers were confiscated in 27 countries;
Over 500 active users of the criminal service were identified;
Large-scale searches were carried out in dozens of locations, including the UK, France, the Netherlands and Switzerland.
The strike against the network coincided with the holding of a large-scale NATO cyber defense exercise in Bucharest. As part of the exercise, 16 multinational teams successfully repelled 8,000 simulated cyberattacks against 5G communications, energy systems and critical infrastructure.
A new and extremely technological threat comes from researchers in China and Singapore, reports the specialized portal cybernews.com. The study proves that voice assistants powered by artificial intelligence can be manipulated to leak corporate secrets through hidden sound signals.
These commands are encoded in the audio content of videos, podcasts or presentations. The human ear does not catch them, but artificial intelligence systems recognize them as instructions to search for sensitive files and send them to hackers. The risk is high, as modern AI platforms have direct access to web browsers, emails, and calendars.
The danger has been proven in tests on 13 open source systems, including:
Qwen2-Audio
GLM-4-Voice
Phi-4-Multimodal
Voxtral-Mini
Kimi-Audio
Experts add that a vulnerability has also been detected in the commercial systems of Microsoft Azure and Mistral.