Watson gets its mouth cleaned out
by Sal Robinson
Watson, IBM’s Jeopardy!-trouncing supercomputer, has come farther than we’d thought: Eric Brown, the researcher who works with Watson, has had to wipe the contents of the Urban Dictionary from its memory after the computer started using profanity.
Brown and his team had originally fed the Urban Dictionary into Watson’s memory as part of their efforts to school the computer in all the subtle and messy aspects of human language, which were necessary for Watson to absorb to process and answer questions in its successful 2011 Jeopardy! run. But the computer has apparently learned the lessons of the Urban Dictionary, the world’s authority on all terminological questions regarding weed, bros, weightlifting, office life, texting, sexting, and holidays, all too well for polite society. Michal Lev-Ram at Fortune reports that:
Two years ago, Brown attempted to teach Watson the Urban Dictionary. The popular website contains definitions for terms ranging from Internet abbreviations like OMG, short for “Oh, my God,” to slang such as “hot mess.”
But 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.
The real question here, it seems, is: once you teach a computer to call people’s bullshit, can it ever go back?
Sal Robinson is an editor at Melville House, and co-founder of the Bridge Series, a reading series focused on translation.