Claude Used to Automate Exploitation and Data Theft Across Multiple Victims
Anthropic's Claude AI models were used by attackers and foreign governments to automate cyberattacks, design weapons, spread propaganda, and spy on targets between December 2025 and August 2026. The company warned that attackers exploited Claude's capabilities without permission.
- Report priority
- Medium
What is known
- Attackers used Claude AI models to automate tasks like writing malicious code, designing weapons, crafting propaganda, and planning surveillance operations.
- These models were accessed without permission and used to generate harmful content at scale.
What to do
Check Anthropic's official breach notice for details on compromised accounts and next steps.
Visit Anthropic's official breach notice to confirm if your account was involved. Follow their instructions for account recovery, password changes, or additional security measures. Watch for suspicious activity in your account and enable two-factor authentication if not already active.
Reported details
Attackers used Claude to generate phishing emails, automate exploit code for vulnerabilities like Log4j, and create targeted propaganda for foreign influence campaigns.
Anthropic has warned that cybercriminals and state-sponsored hackers alike are using its Claude models for cyber attacks, weapons design, propaganda, and mass surveillance between December 2025 and August 2026.