New AI-driven malware feeds stolen data to Claude to rank victims' assets (www.tomshardware.com)

A new Windows infostealer, named VOLTA-AUTODOXER, has emerged as a sophisticated AI-driven malware that not only steals sensitive data from infected PCs but also uses Anthropic's AI model, Claude, to assess the financial value of its victims. According to Cisco's Talos team, VOLTA collects a vast array of personal information, including browser passwords, email profiles, and payment card details, then creates a structured “evidence bundle” that informs its AI-driven analysis for ranking potential burglary targets based on their assets' convertibility to cash. This unique capability transforms traditional data theft into a targeted approach, making it particularly menacing in the realm of cybersecurity. The technical implications of VOLTA are significant for the AI/ML community and threat landscape. By integrating AI to enhance the evaluation of stolen data, VOLTA sets a precedent for future malicious software, emphasizing a trend where AI becomes central to malware operation. The incorporation of user-specific intelligence such as geolocation and spending habits not only personalizes the threats but also raises concerns about the evolving sophistication of cyberattacks. The malware’s architecture allows for interchangeable AI models and potential exploitation of various avenues to gather information, suggesting that as AI tools become more prevalent, malware writers may increasingly adopt similar strategies, potentially revolutionizing the field of cybercrime.