Watson 1: Thinking About Machines That Think

Okay, a while back I posted a link to a story on the New York Times webpage, “Armies of Expensive Lawyers, Replaced by Cheaper Software,” by John Markoff. I said that it was very important.

You may have wondered why. I meant to explain myself quite a while back, but then, Gadhafi and Japan intruded. They were so much more important than any thoughts of mine.

But, now, I’ll get back to my point.

The software that is replacing legions of lawyers is a part of a much, much larger trend—that is, the development and application of technology which can largely automate certain aspects of white-collar work. The products mentioned in the piece allow one lawyer to do the work of many.

And it’s just beginning. Did you watch IBM’s Watson on Jeopardy? That machine’s performance was very much a sign of things to come.

Oh, don’t get me wrong. Watson isn’t going to take over the world. He’s harmless.

That’s because Watson isn’t really intelligent in a human sense. Watson’s showing was impressive. His ability to understand human speech was a technical triumph, and his capacity for seeking commonalities in huge collections of data was astonishing. But, I am not at all certain that he really understood his answers.

For instance, in one particularly amazing exercise (later highlighted on PBS’s NOVA), he was able to answer a question about Keanu Reeves and the movie, The Matrix. The way he did it was to identify “Reeves” as a noun and “movies” (or, actually, “flick”) as a category. He was then able to scan the Internet Movie Database in all its vastness for combinations of “Reeves” and some other qualifiers from the question.

As I say, impressive. But, ultimately, his answer wasn’t complete. I’m not sure he knows what a “movie” is. And I’m certain that his conception of a movie, should he have one, is not in any way human. He knows a “movie” is a “flick,” and that The Matrix is both of those…but he has no conception of what it’s like to see a film. He has no image of being in the warmth of a darkened theater, the air redolent with the scent of popcorn and butter, while your young son and his friends are in the row before yours, and you watch them tense and leap and cheer as villains are defeated and heroes triumph.

So that, I think, should be reassuring to anyone who fears that intelligent machines are taking over the world. Watson is amazing, but he cannot genuinely understand the human experience. It will be many long years before machines can do those things, if they ever do. (Sorry, gang, Skynet and Robby the Robot remain as distant as ever).

That’s the good news.

Now, the bad news. Watson and similar technology will engender a revolution in the workplace.

Let’s look at what Watson does really, really well. He—like all computers—excels at performing a number of small, tedious tasks over, and over, and over again. And, he is really good at shifting oceans of data in search of details and connections, no matter how trivial. And, finally, he obeys rules, inflexibly and tirelessly.

That was why he was so good at answering questions. He looked at millions upon millions of records in a thousand different databases, checking each and every one of them for certain qualities, and doing it all in nanoseconds. A human could never dream of doing the same, nor would any of us want to. It would be exhausting, and, bluntly, boring as hell.

But here’s where things get sticky. Consider what most high-powered, high-paying “professional” jobs entail. Sure, there are moments of insight and creation, but, let’s face it, most of what lawyers, business professionals, research-oriented academics, and even doctors do is pretty much what Watson does. They look at a lot of information and seek connections within an ocean of small, possibly relevant details. They do it according to a fairly limited set of rules. They do it as quickly as they can, and, if possible, do it 24/7 because that means they’re being more “productive.”

In other words, most of our white-collar professions, the jobs we value the most, are made up rote tasks.

And there’s the rub. Watson is far better at rote tasks than we can ever dream of being. His software and circuits are tireless. He is never bored or depressed. No operation is too mundane to excite him. No amount of data is overwhelming.

And better still, he is a perfect employee. He is without ego and need. He will happily work around the clock and back again. He is never on vacation. He never requires a sick day. He has no family to distract him from his tasks. He does not ask for a raise or a bonus. When the time comes to retire him, he demands no pension. You simply toss the body into a dumpster.

No, I’m not saying that Watson’s heirs will take over the corporation. Bill Gates and Tina Brown will not be replaced by a Terminator in the corner office. (At least not yet.) But the inescapable fact of the matter is that Mr. Gates and Ms. Brown will need fewer and fewer people to carry out their commands. A single executive, armed with one Watson, can do the work of a hundred MBAs.

And make no mistake, the world’s corporations know that. Watson is unique today, but he’ll be ubiquitous tomorrow. Even if he never sits on your desk like a PC or Mac, he’ll be available by subscription through the ‘Net. (And isn’t that what Google’s search engines are actually evolving toward?)

Which means that the white-collar professions (including the service professions which were supposed to save us in the postindustrial age) are going to employ fewer and fewer people. Just as the tractor removed the need for legions of field hands, and the industrial robot replaced a million men on assembly lines, so too will Watson and his children allow corporations, government agencies, and just about every other sort of organization to shed thousands of all too human professionals.

That, in turn, means that things are going to change. And big time.

But more about that next time.

Onward and upward.