By Natali Moss
According to the company, the generative AI allowed to use tactics without harmful software, automate lateral movement and scaled social engineering. "The villains are now operating at the speed of machines in different domains, which makes many traditional remedies outdated. This is a revolution in the surgery of malefactors, and it requires a new approach to security," Crowdstrike emphasized.
The report separately mentions the DPRK -related Famus Collima, which has penetrated more than 320 companies during this period, using services with large linguistic models (LLM) to create resumes, fake identification and even answers to real -time interviews. After employment, the self -esteem relied on artificial coding assistants and translation tools to simultaneously perform several remote tasks for the developers, simultaneously stealing intellectual property.
In order to avoid such situations, Crowdstrike recommends advanced person check procedures during recruitment, real -time testing for fakes during interviews, more careful monitoring of remote access activity, as well as constant verification of identification data, endpoints and cloud telemetry. "The era of artificial intelligence has changed how the attackers work, and most protective mechanisms were not created for this reality," experts emphasized.
English
Українська
USA
Français
Deutsch
Italiano
Polski
Čeština
Español
Slovensku
