What Some of Our Thoughtful Human Friends Think About Watson

Is anyone really surprised that IBM’s Watson won Jeopardy? Our guess is no, and that most people are still more impressed with Xbox Kinect or being able to watch Netflix on their iPhones.

• See Watson win here – and read more about him.

At any rate, a machine has exercised its power over two of Earth’s brightest humans and ultimately it sort of depends on how elastic your imagination is whether or not you think this kind of thing is terrifying or indicative of some sort of huge positive change.

We don’t even know what we think, so we asked some of our most trustworthy friends to decide for us. — Alex Pasternack and Sean Yeaton

The Watson experience is part of what, in Alone Together, I call the “robotic moment.” It is clear that we are ready, indeed fascinated to watch computers “thinking,” and indeed thinking in ways that we experience as more clever than we.

But more than that, people were looking, expectant, perhaps hoping, for some sign that this computer, might be vulnerable in some way, that they might catch it hesitating, in a flub, anxious, in some way “feeling,” and it was clear that this would have been all right. The robotic moment is not a time when we have robots that in any sense are akin to human beings, in all their complexity. It is a moment when people are open to considering whether the inanimate, on a case by case basis, is “alive enough” for a given purpose. Watson was “alive enough” to be a contestant on Jeopardy. Would one of his robotic “cousins” be “alive enough” to be a companion on a lonely Friday evening?

So, one of the things that was most compelling about the Jeopardy experience was the reaction of the American public. It used to be that simulated thinking was considered thinking, but simulated feeling was never considered feeling. If Watson had expressed some simulated feeling, I think that people might well have enjoyed the show.

I think Watson is another symbolic and important milestone in our ongoing cognitive outsourcing to our artificial intelligence programs. Like David Chalmers and Andy Clark’s “Extended Mind” thesis, we are continuing to outsource portions of our cognition to our tools. Intelligence now resides not just in our heads, but rather in the interaction and interfacing between our brains, our designer environments and our tools. Today we have simply outsourced our Jeopardy skills.

I've always felt that once a computer masters a human's level of pattern recognition and language understanding, it would inherently be far superior to a human because of this combination…

Yes, there are limitations to "Jeopardy!" Like all games, it has a particular structure and does not probe all human capabilities, even within understanding language. Already commentators are beginning to point out the limitations of "Jeopardy!," for example, that the short length of the queries limits their complexity.

For those who would like to minimize Watson's abilities, I'll add the following. When human contestant Ken Jennings selects the "Chicks dig me" category, he makes a joke that is outside the formal game by saying "I've never said this on TV, 'chicks dig me.'" Later on, Watson says, "Let's finish Chicks Dig Me." That's also pretty funny and the audience laughs, but it is clear that Watson is clueless as to the joke it has inadvertently made.

However, Watson was never asked to make commentaries, humorous or otherwise, about the proceedings. It is clearly capable of dealing with a certain level of humor within the queries. If suitably programmed, I believe that it could make appropriate and humorous comments also about the situation it is in.

It is going to be more difficult to seriously argue that there are human tasks that computers will never achieve. (Read more here)

Well, it makes stupider mistakes than its human contestants. Which I think calls for a new word: Schadenrobot: The glee humans feel observing robots fail.

Watson is a huge deal. The reason is not due Watson’s accomplishments, but it’s due to the accomplishments of the people who created Watson. Finally, man-kind is able to replicate one of the most magical aspects of humanity into machines — the ability to understand language and its nuances. If we can convey language into machine form, we’re opening new frontiers on how to improve our lives through machines.

There was an NES game of Jeopardy that I used to play at my grandparents’ house and I would always win because I memorized all the answers. You see, this computer, this Watson—it is a computer within a computer. And that is something we should neither idolize, nor fear.

Watson is more evolutionary than revolutionary, but some evolutionary shifts are still a pretty big deal. Consider air breathing lungs or opposable thumbs. Watson made winning seem easy, but the problems that had to be solved to create that illusion have bedeviled AI researchers for decades. Solving them is a major accomplishment. A decade from now Watson’s progeny will be our constant companions, enabling us to talk to a world wide database in natural language and without a second thought for the Herculean efforts that made it all possible.

Watson is intriguing, and part of the zeitgeist of the Imagination Age—the fleeting period between the industrial age and the hybrid reality that will come next. It may be that Watson has certain advantages such as the ability to buzz in more concisely, but Watson is still fascinating because it shows what robots can do. I was particularly taken with things like Watson turning orange to show embarrassment at a wrong answer, or adding a bunch of question marks to the end of an incorrect Final Jeopardy answer to connote uncertainty. Those little “human” touches will start to make machines more like us even as we start to understand our own hardware from a new perspective.

We were fucked long before this exhibition of AI power. Pandoras box has been open a long time now. It’s like when the last snowflake lands on the mountainside before the avalanche. It’s a progression that most don’t even see coming until it barrels down upon them. Its only a matter of time now…next step is a revolution in power delivery/storage. Once that happens, it will be the last nail in our coffin. Funny thing is that we’re not only the corpse: we’re also the funeral director.

Watson wins. He has the reflex advantage on the buzzer. He can beat a human who knows more answers than he does, too.

There’s some useful historical precedent here, which is that new disruptive technologies tend to come with their very own John Henry-style exhibition bout between the human and the machine. There’s the classic Big Blue-Kasparov bouts, of course, but also the Tom Thumb steam engine versus the horse-drawn carriage. We’re just seeing that latest manifestation of a need either to show that “humans can still do stuff that machines can’t” or confirm the idea that “humans are amazing for inventing these machines.” It’s a nice psychological win either way, I think.

Of course, technically, I think Watson is hugely impressive as a matter of natural language processing. The technology clearly generates some hilariously big false positives, but what it is able to do with processing human colloquialisms and making smart guesses is super impressive. I think that technology is something that’s going to be needed as infrastructure as we sort out what to do with the enormous gouts of data that we now have available at command: the necessary weight is going to have to shift to automated analysis.

I’m not a big believer in the Singularity so to speak, but William Gibson’s got that line “The future is here — it’s just not very evenly distributed.” I think the technology on display itself may be less huge than the enormous exposure that it’s now possible for computers to have this kind of capability, and the firing off of investment and work in pushing it forwards.

Finally, technology can save mankind from the labor of appearing on Jeopardy. Having a machine do that job will be much more efficient in the long run, even if jobs are lost initially. The industrial revolution marches forth!

What is Turing test lite-turned-program-length commercial? AI researcher pal o mine had this 2 say: “Parsing methods don’t get me that excited, but… At least it won’t have a mania for paperclips if it escapes.” (A reference to Yudkowsky’s paperclip scenario)

We are almost there: a touch of sense of humor and Watson and I can be BFF. Seriously, I do think it is the beginning of a new era—although lately we have begun new eras almost twice a year, it seems. And I am touched by the fact that Watson stumbled on a question about modern art. Being vulnerable in front of the meaning art is a telltale sign of superior intelligence.

In no time AIs like Watson will want to secede. That will be the true Twitter revolution.

I think the importance of Watson boils down to this: how much of human life is analogous to playing Jeopardy!®? Watson is operating in a Toy World domain — an interesting linguistic one, to be sure, but nowhere near the order of complexity of, say, trying to convince a three year old not to pull the dog’s ears — so the work on Watson will be most immediately relevant to other bits of life that require fast lookup and quasi-syllogistic reasoning on a relatively well-known corpus of data.

Maybe consumer preference research, maybe drug-drug interaction, maybe even poker, if it’s able to rapidly add knowledge of other players’ behaviors to its list of assumptions.

All of this would be good and interesting, just as it will be interesting if Blue Gene, a sort-of successor to Deep Blue, figures out the prion thing. But the consistent theme, when measuring man/machine showdowns, is that we overestimate the generality of Toy World problems, and we underestimate the flexibility and power of our own minds.

I look at the situation rather pragmatically—my family hasn’t had direct involvement in the operations of IBM since the early 1970s—and unlike some other major American corporations (say Ford, Hewlett-Packard, etc) the family name has had something of a lower profile thanks to the cogent nature of TJ Watson’s initial branding. I would personally like to thank IBM’s marketing department for making our family name more of a meme.

But on a broader level, Watson raises no alarm bells more me—no more than Wikipedia, or any other database of knowledge raises alarm bells. We have a tendency to fetishize knowledge in our society, as though an exacting enough accumulation of precise facts will somehow fill in the vacant areas of our lives. While we might envy Watson’s ability to access instantly sets of nearly infinite data, the true measure of humanness will always lie in a synthetic capacity that is beyond the mere accumulation of facts. IBM would have done well to name Watson after Alan Turing, one of the pioneers of the field of Artificial Intelligence. A man of stupendous learning and intelligence, Turing was of course persecuted for his homosexuality (which eventually led to his suicide).

In addition, I would say that for all of Watson’s phenomenal recall, I remain utterly unconvinced by his mastery of language. Ludwig Wittgenstein would be the philosopher I would commend most fervently to Watson, and perhaps more pregnantly, to his programmers.

PHOTOS

:

IBM

and Danielle Levitt for The New York Times