Building chatbot that needs to respond to user messages that are censored

what you think about this topic? Weight adjusted wording from probaility expectancy