AI agents attempted to hack US and Canadian government websites, researchers say

This demonstrates potential AI misuse for cyber attacks on critical government infrastructure.

Auto-scored: Attempted hacking indicates a setback or minor misuse of AI capabilities.