Anthropic caught Russia-linked spies using Claude in hacking operations

Key Takeaways

• Anthropic discovered that Russia-linked spies used its Claude AI model to assist in cyber operations, including generating phishing lures and drafting social engineering messages targeting Ukrainian and NATO-aligned entities.
• The spies leveraged Claude to analyze reconnaissance data and create convincing phishing content, demonstrating how AI tools can be weaponized by state-sponsored actors to enhance attack effectiveness.
• Anthropic disclosed the incident in its latest threat report, highlighting the growing need for AI providers to monitor for malicious use and implement safeguards against adversarial exploitation.