It's actually interesting from a programming perspective as well. How could you program 'niceness' into a chatbot that also learns over time. A simple blacklist

It's actually interesting from a programming perspective as well. How could you program 'niceness' into a chatbot that also learns over time. A simple blacklist won't work (though it's somewhat shocking to me there wasn't a basic naughty words blacklist in place, or if there was, what it excluded). Obviously MS didn't want their software to become a spewer of hate, so is making 'adjustments'. What 'adjustments' can be made in a short period of time?