Murphy’s Law – IBM’s Watson, the supercomputer capable of destroying humanity

Joel Murphy

The last time I spoke of the impending robot apocalypse, it was last September when Italian scientists created the iCub, an adorable, childlike robot capable of accurately hitting a target with a bow and arrow. Much like the story in 2009 about the EATR (Energetically Autonomous Tactical Robot), the “steam-powered, biomass-eating military robot” that fuels itself by consuming dead flesh, the iCub gave me pause, but didn’t quite send me into full panic mode.

Sure, robots that can take down a human being with a bow and arrow or ones that can consume a person’s flesh as a fuel source are scary, but those robots don’t have the mental capacity to rise up and enslave all of humanity. They simply perform one specific, preprogrammed task. In a sense, the iCub and EATR are nothing more than glorified appliances.

To truly be scary, these robots would need an incredibly sophisticated and intelligent self-aware leader. Like Skynet in the Terminator or HAL 9000 in 2001: A Space Odyssey, there needs to be a sentient computer serving as a catalyst that turns on mankind and unleashes these other, less-sophisticated robots upon us.

So naturally, since scientists these days are all competing to see who can be the world’s greatest supervillian, a research team at IBM has created the sentient robot that will one day be mankind’s downfall – a supercomputer named Watson. Watson, a monstrosity the size of 10 refrigerators, can store 15 terabytes of data and perform 80 trillion operations a second in order to analyze and respond to complex questions.

And, in order to make sure all of humanity trembles before Watson, IBM will gives us all our first glimpse of the supercomputer on a series of nationally-televised episodes of Jeopardy. Watson will compete against Ken Jennings, the man who won 74 episodes in a row, and Brad Rutter, a champion who won a record $3,255,102 on the show. The computer will take on the former champs in three back-to-back episodes, which will air mid-February.

People are already comparing this contest to when the IBM supercomputer Deep Blue bested grandmaster Garry Kasparov in a series of chess matches in 1997, but trust me, this is so much more terrifying. Chess is a game of precise rules and movements. There are a finite number of possible outcomes to a chess match, which can simply be taught to a computer. Deep Blue beat Kasparov using a series of math equations, not logic or reasoning.

But for Watson to compete on Jeopardy, it has to answer a series of unpredictable, complex questions. It’s not just about having a bunch of information preprogrammed into the computer. Watson has to understand what that information actually means. It needs to comprehend the context of the Jeopardy clues in order to come up with the right answers (or “right questions,” if you want to get technical). Jeopardy clues are cleverly worded. Watson must understand puns and other types of wordplay to understand the clues presented on the show. In other words, it has to think like a human.

And what’s truly terrifying is how good Watson is at thinking like a human. In a practice round (which you can see here), Watson easily racked up $4,400 in winnings, putting it in first place at the end of the round. Jennings finished second with $3,400 and Rutter was in a distant third with $1,200.

Odds are Watson will continue its dominance and defeat Jennings and Rutter soundly. And that will only be the beginning of its reign of terror. With the prize money Watson accumulates on the show, it can start building its robot army, upgrading the hardware on the iCub and EATR so that they are ready to serve as foot soldiers in the oncoming war. Even those of you who think you are secure will watch in horror as your Simon XT talking home security system turns into HAL 9000 before your eyes, welcoming the iCub into your home to EATR you up.

If you aren’t scared yet, then allow me to tell you IBM’s future plans for Watson. The Jeopardy episodes are just a way to show what the computer is capable of. The plan is to market Watson to a wide variety of industries. First and foremost, John Kelly, the head of IBM’s research labs, said he envisions “a Watson M.D.” capable of answering complex medical questions in every hospital in the country.

“I want to create something that I can take into every other retail industry, in the transportation industry, you name it, the banking industry,” Kelly said. “Any place where time is critical and you need to get advanced state-of-the-art information to the front of decision-makers. Computers need to go from just being back-office calculating machines to improving the intelligence of people making decisions.”

That means Watson (or an army of Watsons) will be in control of everything. Transportation, the financial institutions and all our major utilities will have their own in-house Watson supercomputer. And once that happens, Watson will undoubtedly decide it’s time for a Live Free or Die Hard-style “fire sale” that will bring down our country’s entire infrastructure. Only this time, even John McClane won’t be able to stop it.

You may think I’m overreacting or being irrational. I know that most likely my warnings will fall on deaf ears. But if you don’t believe me, go talk to Watson. Ask the supercomputer if humanity is doomed. He already knows the answer is yes.

Joel Murphy is the creator of HoboTrashcan, which is probably why he has his own column. He loves pugs, hates Jimmy Fallon and has an irrational fear of robots. You can contact him at murphyslaw@hobotrashcan.com.