It would be good to provide a way for content creators to provide feedback on the classification their text receives. It could be the be the case that some text has been flagged as racist when in fact it is not. This feedback would then need to be added into the AI model and help the AI adjust over time.