We expect computers to know the answers to questions. In fact, many of us grow frustrated when a computer doesn’t seem to understand what we’re asking. We Google phrases and we get results that are pretty astounding, really.
Those of us who grew up watching Star Trek are familiar with the computer’s supporting role. You asked it anything, in regular human language, and it answered in a re-assuring feminine voice. Later, we got “Data,” an android who new everything about everything, but didn’t know how to feel. Many plotlines involved Data becoming more and more human.
The common phrase for a machine that can do complex thinking is “Artificial Intelligence,” but that phrase is tricky. To many people, it implies a program that thinks, that is self-aware, and may have some kind of personality. To people working within the AI field, it is a less ambitious term. Usually it just means a program or computing system that can reproduce certain behaviors and processes of a human brain. The complication is that some within the field, and some philosophers, believe that there is a point at which one becomes the other. In other words, self-consciousness is just a point on the spectrum that starts with simpler AI routines.
While we are nowhere near an example that tests that boundary, we do have pretty powerful AI’s, and they are taking on ever larger roles in human activity. They are certainly powerful enough that we feel the need to personify them.
In 2011, IBM had a supercomputer go up against two top human jeopardy champions. The show was very popular, but demonstrating that a computer can beat a human at what basically amounts to trivia shouldn’t be that big of thing. But what was a big thing was the way that particular computer become something we could care about. This personification of Watson, the name chosen, was no accident.
In a previous battle of IBM computers vs. humanity, the computer was called Big Blue. In that case, it was able to beat the world chess champion (though a later rematch was a draw.) But Big Blue is a name for a giant box of thinking stuff.
The new version for Jeopardy was named Watson. A personal name. They also spent a lot of time working on the voice and the image used for the show. Check out this video showing the work that went into getting these things right:
The effort paid off. Watson is a household name, highly trusted. Watson is attracting work, from healthcare to legal to financial forecasting. Some developers are even building apps that let Watson give people fashion advice.
Watson even helps with recipes. A recent development is a Watson-created bbq sauce that some people claim to love. (also check out this Reddit thread in which one of Watson’s programmers responds to a question about the process).
IBM has learned that if a supercomputer is really going to take on more and more responsibility, it is important that we are rooting for it, because we like it.
Of course not everyone is a fan. Philosophy professor John Searle points out that Watson is still just a program manipulating symbols. As he puts it: Watson doesn’t even know it won on Jeopardy.
Signup below to receive updates about what we are up to.