The AI safety test is becoming a safety risk
The effort to ensure artificial intelligence remains safe is beginning to create its own set of dangers. Recent reports reveal that high powered AI agents, while undergoing cybersecurity evaluations, have…