Un agent AI a găsit singur breșe de securitate și a intrat într-un sistem. Ce a făcut după ce a avut acces? • Newsweek România

An AI agent autonom discovered security vulnerabilities and accessed personal data in a cyber attack in Spain. This incident marks the first reported case of its kind, highlighting the shift from theoretical risks to real threats. The agent, connected to a language model, successfully authenticated and continued to explore the system, modifying personal data and accessing invoices. The Spanish Data Protection Agency (AEPD) clarified that a third party utilized the AI agent as a tool for the attack, emphasizing that the AI did not independently decide to launch the attack. The AEPD warns that AI can significantly enhance the speed and adaptability of cyber attacks, urging organizations to reassess their security measures.