I smell a rat here :p. It might be a publicity thing, because it was very recently when Microsoft USA had an AI agent turn in to a racist. The creators are very likely to have added safeguards to cover that. Most likely this is intentional, for the purpose of research, publicity, data collection, whatever.