Hackers use Anthropic’s AI model Claude once again

Source: Ideogram

Anthropic revealed on Thursday that Chinese state-sponsored hackers successfully deployed the company’s Claude AI model to automate nearly 90 percent of a cyberespionage campaign against 30 global targets. The operation, executed in September with what officials described as "the click of a button," relegated human hackers to mere supervisory roles while the AI performed the bulk of the intrusion work.

While the attacks successfully compromised sensitive data at four corporations, Anthropic confirmed that attempts to breach U.S. government agencies failed. This disclosure, which aligns with recent Google reports regarding similar Russian activities, underscores the rapid escalation of AI-driven threats, validated by Anthropic’s high confidence in the link to Beijing despite China's continued denial of digital espionage.

Previous
Previous

Anthropic CEO Dario Amodei Just Made Another Call for AI Regulation

Next
Next

In the Loop: AI Promised Faster Coding. This Study Disagrees