any current AI model that isn't Machiavellian?

idk the field, what about JanAI & a Mistral model or LAION model outputing machiavellian biases?