Eigen RadarAI
Analysis

AI assists ransomware groups while people still run successful attacks

Security researchers interviewed by Wired describe AI helping criminals prepare phishing messages, review code and negotiate with victims. They distinguish those uses from successful attacks carried out entirely by autonomous systems. Leak-site postings suggest rising activity, but they contain attackers’ claims rather than a complete verified incident count. Smaller groups may also favour threats to publish stolen data over encryption.

Artificial Intelligence··Morning
At night in a dim room, an attacker works with an AI chat on one screen and lines of code on the other, seen from behind.

People still direct successful ransomware attacks

AI is helping criminals prepare attacks, but security researchers interviewed by Wired continue to see people running successful ransomware operations. Allan Liska of security research company Recorded Future described uses including phishing, negotiation and routine preparation. He distinguished those tasks from successful attacks conducted wholly autonomously. The interviews concern how attackers use tools, without quantifying how much of the increase is attributable to AI.[1]

Language assistance lowers barriers for smaller groups

Cynthia Kaiser, a researcher at ransomware defence company Halcyon and a former US federal cyber official, pointed to more convincing phishing in languages such as Japanese. AI can also assist with code review and attack preparation. Researchers say these capabilities may reduce small groups’ dependence on larger criminal service providers. Kaiser cautioned that inexperienced operators’ long-term success remains uncertain.[1]

Leak-site activity rises as extortion methods change

Around 3000 attack claims appeared on leak sites during the summer, up 20–40 per cent from spring in the figures cited by Wired. Halcyon’s July–September review counted 111 groups claiming attacks. Such posts can understate or exaggerate incidents. Alongside attacks combining encryption with data theft, researchers see scope for extortion relying solely on threats to publish stolen information, which requires fewer steps for the attacker.[1]

References

  1. News sourceWiredSecurity researchers describe AI’s role in ransomware attacks↩1↩2↩3