Anthropic Says Russia-Linked Hackers Used Claude to Target More Than 20 Organizations
Summary
Anthropic says it disrupted a Russia-linked campaign that used Claude against more than 20 organizations, and describes other groups using AI for credential theft, data theft, vulnerability research and attacks on political and media targets.
Key points
- Anthropic says a campaign aligned with Russia-linked Midnight Blizzard targeted government, intelligence, diplomatic and defense organizations, including Ukrainian entities and drone supply-chain organizations.
- Anthropic reports that the group accessed mailboxes at two drone-component manufacturers, stole a proprietary drone-vision software development kit and used Claude to analyze the system.
- The report says the group compromised hotel Wi-Fi providers, changed DNS records to redirect travelers, and used Claude to modify hacking artifacts after security products detected them.
- Suspected ShinyHunters affiliates reportedly used AI to find credentials, map systems and steal data for extortion; one operator allegedly escalated from a stolen developer token to cloud administrator access in about three hours.
- Anthropic describes a Chinese-speaking group conducting vulnerability research against a major security product and a French-speaking hacktivist targeting European political parties, media organizations and think tanks.
- Anthropic says it disrupted the activity, strengthened safeguards and shared intelligence with authorities and industry partners; the company’s report does not provide broader figures on the scale of detected misuse.
- The article notes that phishing, stolen credentials, exposed services and software flaws remained central to successful attacks, despite the use of AI.
Article Details
- Event Type
- Anthropic detected and disrupted misuse of Claude in cyber operations, including a Russia-linked espionage campaign and activity by criminal groups and individual hacktivists.
- Impact
- The campaign targeted more than 20 government, intelligence, diplomatic, and defense organizations. The Russia-linked group stole a proprietary software development kit for a drone vision system after accessing mailboxes at two drone-component manufacturers, and used Claude to analyze the system and adapt detected artifacts. ShinyHunters affiliates used AI to find credentials and steal data; in one case, an operator reportedly escalated from a stolen developer token to administrative access in a victim's cloud environment in about three hours. Anthropic said it disrupted the activity and strengthened safeguards.
People
Threat Actors
Cozy BearName identified in the article as an alternate name for Midnight Blizzard.ICE RELICName identified in the article as an alternate name for Midnight Blizzard.Midnight BlizzardRussia-linked espionage group whose campaign used Claude and targeted government, diplomatic, defense, and drone-related organizations.ShinyHuntersCybercriminal group whose suspected affiliates used AI to find credentials, map systems, and steal data for extortion.Storm-2945Sub-cluster of Midnight Blizzard linked by Microsoft's investigation to the described activity.
Vendors
AnthropicAnthropic said it detected and disrupted a Russia-linked espionage campaign that used Claude to target more than 20 government, intelligence, diplomatic, and defense organizations.GoogleAgranovich, who went on to found Meta’s threat-disruption team and now works on adversarial security at Google, cautioned that “some press coverage is going to frame this report as ‘Claude was used to [do bad thing]’MetaAgranovich, who went on to found Meta’s threat-disruption team and now works on adversarial security at Google, cautioned that “some press coverage is going to frame this report as ‘Claude was used to [do bad thing]’MicrosoftAnthropic cited Microsoft’s investigation linking the activity to Storm-2945, a sub-cluster of Midnight Blizzard.OpenAIBut unlike similar reporting from the company’s rival OpenAI, Anthropic includes in-depth analysis of several campaigns and abuse types, alongside indicators of compromise (IOCs) that can be used by security teams.
Products
Countries
Industries
DefenseAnthropic said it detected and disrupted a Russia-linked espionage campaign that used Claude to target more than 20 government, intelligence, diplomatic, and defense organizations.GovernmentAnthropic said it detected and disrupted a Russia-linked espionage campaign that used Claude to target more than 20 government, intelligence, diplomatic, and defense organizations.intelligenceAnthropic said it detected and disrupted a Russia-linked espionage campaign that used Claude to target more than 20 government, intelligence, diplomatic, and defense organizations.MediaAnthropic also described a French-speaking hacktivist who used Claude in attacks on several European political parties, media organizations and think tanks.