Can AI Lie? Detecting Deception with Chain-of-Thought Monitoring