Emma and Dewey are now much better at dealing with those few odd folk that insist on using profanity. A few modifications to Square Bear's shutup.aiml and now conversation ends when profanity is used. At least until the user apologizes. Just like at the reference desks with real, human librarians and patrons, it's nice that our bots don't have to take abuse.
[In Pandorabots' SuperBot 2.0 most of the problem words are reduced to PROFANITY or INSULT. If you're not running SuperBot, it's still pretty easy to reduce all of them to one or two patterns, then alter shutup.aiml accordingly. I was tempted to add responses like "recording client barcode and location," but we all know what would happen if I did. It's probably best just to make them say they're sorry.]