https://vinarise.com/en/topics/ai-safety Making AI systems do what we intend, and knowing when they don't. en