AI-Powered PowerShell Script Used in New Cyberattack
Cybercriminal groups are refining their methods by using AI to generate malicious code. A new report from Proofpoint reveals how the threat actor "TA547" is using AI-generated PowerShell scripts in a targeted campaign against German companies.
The Threat Actor "TA547" Uses Advanced AI to Target German Companies
Security firm Proofpoint has discovered that the cybercriminal group "TA547" has launched a new email campaign targeting organizations in Germany. The campaign exploits malware called Rhadamanthys and uses emails that appear to come from the well-known German retail company Metro. These emails claim to contain invoice information but actually contain a password-protected zip file. The file hides an .lnk file that, when activated, triggers a PowerShell script to run and decode a Base64-encoded version of Rhadamanthys.
AI-Generated Code Identified in the Script
The most notable aspect of this attack is that the PowerShell script used appears to have been generated by an AI. Proofpoint points out that the script contains detailed comments for each code component, a characteristic typical of code created by advanced language models such as ChatGPT, Gemini, or Copilot. This phenomenon raises the question of the extent to which cybercriminals may come to use AI technology to automate and improve the execution of their attacks.
The Same Defenses Apply Against AI-Generated Threats
Despite the potential involvement of AI in creating malicious code, Proofpoint emphasizes that the defense against these threats doesn't change. The fundamental principles of cybersecurity, such as caution with unknown email attachments and password protection, remain just as crucial regardless of the threat's origin, whether it's manually coded or AI-generated.