IBM’s Watson Cursed Like a Sailor After Being Taught the Urban Dictionary
It’s a fact that Watson, IBM’s massive AI project, is smarter than the average human. I mean, it kicked Ken Jennings’ ass on Jeopardy that one time. “Smart,” in that respect, meant the ability to pull knowledge from terabytes worth of Wikipedia data based on verbal clues.
But Ken Jennings (and you and me) has Watson beat in one measure of intelligence: human language. Once that fact no longer holds true, well, we’re all in a hell of a lot of trouble.
Nevertheless, IBM is trying to improve Watson’s human language prowess. And to do that, Watson needs to understand how humans talk – how they really talk. I’m talking about slang, of course. People simply don’t realize just how complicated human language really is. Teaching Watson proper and direct English is nowhere near good enough to turn it into a fully functional conversation partner. I mean, how the hell is it going to know how to respond to YOLO?
So, to work on that slang element of human language, IBM researchers decided to teach Watson the Urban Dictionary – you know, the online database of anything and everything human begins say – from the inane to the foul.
Apparently, this led to a problem. Watson developed a mouth like a sailor. From Fortune:
Watson couldn’t distinguish between polite language and profanity — which the Urban Dictionary is full of. Watson picked up some bad habits from reading Wikipedia as well. In tests it even used the word “bullshit” in an answer to a researcher’s query.
Ultimately, Brown’s 35-person team developed a filter to keep Watson from swearing and scraped the Urban Dictionary from its memory. But the trial proves just how thorny it will be to get artificial intelligence to communicate naturally. Brown is now training Watson as a diagnostic tool for hospitals. No knowledge of OMG required.
Suck it, Trebek.[Fortune via The Atlantic]