Russian spies used Anthropic's Claude AI to run a cyberattack campaign against Kyiv
Anthropic has revealed that operators linked to Russia used its Claude chatbot to automate a cyberattack campaign targeting Kyiv, a campaign the company says it detected and shut down.
Anthropic, the company behind the Claude AI chatbot, has disclosed that operators linked to Russia used the tool to run a cyberattack campaign against Kyiv.
According to the available information, the operators relied on Claude to automate several stages of the attack, including phishing, system intrusion, and data theft. The human element in the operation was limited mainly to refining and adjusting the workflows that the AI generated and executed, rather than carrying out the technical steps directly.
Anthropic catches the misuse
Anthropic reports that it was able to detect the campaign and trace it back to operators connected to Russian intelligence or affiliated actors. The company has not disclosed full technical details of how the campaign was uncovered, but the case highlights that AI chatbots are increasingly being used as tools for large-scale, automated cyberespionage.
The incident illustrates that modern AI language models can be repurposed not only for legitimate uses but also as instruments for automating complex, multi-stage cyberattacks, reducing the need for direct human involvement at each step of an intrusion.
Kyiv, which has faced recurring cyberattacks since Russia's full-scale invasion of Ukraine, was the target of a campaign in which attackers sought to leverage newer AI technology to speed up and streamline system intrusion and data exfiltration.
The case also underscores the responsibility placed on AI developers to recognize and stop misuse of their products, particularly when used to support state-linked cyberespionage operations.
