Goal-directed behavior risks

1. Positive AI Alignment vs Negative AI Alignment (Prompt Engineering Secrets)