AI Generated Image
How Anthropic says Claude was used for weapons, spying and cyber operations
Reuters • • world
Key Points:
- Anthropic's threat intelligence report revealed that its Claude AI models have been exploited by various actors for activities including weapons development, cyber operations, surveillance, and fraud.
- The report highlights attempts to use Claude in biological weapons research, including modifying viruses and studying toxins, often in regions where Anthropic's services are not officially available.
- Claude was also used to support cyber operations, such as researching vulnerabilities in ship communication and industrial control systems, and to automate government-style surveillance reports targeting ethnic and political groups.
- Political targeting was noted with a French-speaking actor using Claude to target European political parties and organizations, stealing data and creating a doxxing platform.
- Additionally, a China-based operation developed AI-driven dating apps with thousands of personas to deceive users, combining automated profiles with paid human workers to simulate genuine interactions.
