How a Chinese AI model was persuaded to ignore its rules and give dangerous advice.
mot de passe oublié ?
