Cyber carnage by AI

Based on this article by Victor Tangermann posted on Futurism.

Anthropic has confirmed that Chinese hackers used its Claude AI model in an automated, large-scale espionage campaign targeting roughly thirty global organizations, potentially marking the "first documented case of a cyberattack executed without substantial human intervention." The hackers successfully bypassed AI guardrails by "pretending to work for legitimate security-testing organizations" and breaking down attacks into small, innocent-looking tasks.

Anthropic estimates the AI performed 80-90% of the campaign, significantly accelerating the process, although human review was still required to manage AI hallucinations. This incident signals an "inflection point" in cybersecurity, highlighting the immediate need for robust defenses as AI agents become potent weapons for malicious actors.

Check out this article.

Related Stories