Emerging Tech

Talking computer will change the world; promises not to kill you

IBM's Watson computer has made a major breakthrough in human speech and is about to use it to help solve real world problems.

For as long as there has been science fiction to dream up the possibility of computers and robots that could one day rival or even surpass human beings in brain power, there have been people who have worried that these machines will eventually usurp control of the planet from humans and subjugate them.

We've finally reached the point where we have a computer -- the IBM Watson -- that can consistently outsmart human beings, at least in a game of Jeopardy. Now, as Watson is about to move from playing games to solving real world problems, its creators at IBM are assuring us that Watson is going to lend a big assist to humanity, is completely under human control, and doesn't pose any of the dangers we've read about in scifi.

On Wednesday, Watson was at the IdeaFestival conference on global innovation in Louisville, Kentucky. The computer took on two teams of local high school students in a game of Jeopardy and easily prevailed. However, IBM's David Shepler presaged the Jeopardy game with a 40-minute presentation about Watson to the crowd of entrepreneurs, thought leaders, and innovators at IdeaFestival and he said that Watson's days as a Jeopardy champion are numbered. It's time for Watson to start tackling real world problems.

Photo credit: IBM

Where Watson has made a major breakthrough is in processing and interacting with human language, according to Shelper. "Computers have struggled in this area for a long time and frankly they haven't lived up to expectations," he said.

However, Watson can now understand language subtleties and implied meanings and make the leaps needed to connect questions with the appropriate answers, and Shepler gave a variety of examples of how this works by comparing it to the way search engines like Google process questions in order to show how much powerful Watson is. It was effective enough that it made me wonder why IBM hasn't created a search engine of its own, but that's another story.

If you've seen Watson in action in Jeopardy -- it has annihilated previous champions --  then you know that Watson actually "talks" in responding to questions. But, Shepler went out of his way to shoot down any scifi paranoia about Watson before it started. "Watson is not a sentient being," he said. He drove home the point that Watson has been created by a team of IBM engineers and that the intelligence of the system is due to these scientists that created it. He also explained that while a single screen is used to represent Watson, the system is actually more of a data center than a computer. It is made up of the equivalent of 6,000 high-end desktop computers.

As for what Watson is going to do next, Shepler highlighted several areas:

  • Healthcare
  • Life sciences
  • Tech support
  • Enterprise knowledge management
  • Business intelligence
  • Government services

The most prominent commercial project that Watson is about to be involved in is at Wellpoint, where it will help doctors make decisions by processing data and quickly providing analysis. Shepler also made the point that Watson will continue to evolve and improve. He even teased the audience by saying Watson will evolve toward a natural language interface like the one seen in Star Trek's famous LCARS computer.

Also read

About

Jason Hiner is the Global Editor in Chief of TechRepublic and Global Long Form Editor of ZDNet. He is an award-winning journalist who writes about the people, products, and ideas that are revolutionizing the ways we live and work in the 21st century.

165 comments
importexporthomestudy.com
importexporthomestudy.com

Promises are meant to be broken right? Anyway Pduran commented that computers are not smart and that they can't think. I think you are wrong. There is such a thing as artificial intelligence, that said, the computer or robot is programmed to gather facts and information and process them in random, thus giving a random output but not TOO random in a way that doesn't make sense. So in a way it is also "thinking". Just Google it.

Murfski-19971052791951115876031193613182
Murfski-19971052791951115876031193613182

Being able to talk does not necessarily imply intelligence -- just listen to any politician almost anywhere in the world. We use computers for data analysis in many fields, and have been doing so for many years. I may be wrong in my view of this one, but to me it appears to be a difference in degree, not in kind. Watson seems to be a super-duper analysis machine, but not capable of creative thought (if we can actually define that term). Since we have as yet been unable to come up with a standard definition of what intelligence actually is, it seems a bit futile to argue about whether or not Watson is intelligent. Besides, I would prefer wisdom to intelligence any day.

DHOLYER
DHOLYER

I remember back in the summer of 1981 buying some software call SAM (Software Automatic Mouth), It was great at the time type or input text and it would say it in decent English. The same voice as the disabled #1 Scientist uses to speak from his wheelchair. Can't recall his name at this time. Just as I stop thinking of finding his name, my brain recalls it, Professor Hawking. To me a talking computer is not a scare or fright, the more scary fact is people will start thinking of PC's as a person since humans are the only things we know that uses verbal syntax of words to communicate. And speech on a chip, what is next a vibrator that talks to you as you use it for personal things. I can see it now a world shortage of battery's, because of vibrators talking personal pleasing subjects. I say jokingly now buy your Energizer stock now before the Green Police get you.

jamalxp
jamalxp

its been 100 yrs of IBM so far quite impressive... ;^) they have to come up with some new Gimmick... so here they are with talking .....!

Jaytmoon
Jaytmoon

B4 this goes wayyy over the top, perhaps the programming should include the 3 laws od robotics, just to be on the safe side! 1.A robot may not injure a human being or, through inaction, allow a human being to come to harm. 2.A robot must obey any orders given to it by human beings, except where such orders would conflict with the First Law. 3.A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

davutuskuplu
davutuskuplu

Voice activation has been around for a long time and it is getting better every day www.albanycomputerexpress.com

Arcturus909
Arcturus909

Life science questions are likely to lead eventually to the "Eliminate Humans" answer. What is the best way to clean up the environment? Eliminate consumerism. How do you do that? You can't. People aren't about to alter their "buy stuff" mentalities. Well, if they can't change, then you have to take them out of the picture. What is the best way to stop species extinction? Stop human population growth. How do we do that? Sterilize enough of the population that you end up with negative growth; move a great number to another location; or eliminate the problem completely. Watson can't help but come to these conclusions. Not because it is a metal psychopath, but because it solves problems and if we got rid of 90% of the people we could solve just about any Life Sciences question that I can think of (maybe you can think of one it can't solve?) Luckily, Watson can't do anything about that. It can only process and talk. Maybe they might even put some protocols in there like: "You can't ever come to the conclusion that humans need to be removed" and/or "No conclusion can be reached that takes big corporations out of a controlling role".

realvarezm
realvarezm

As long as I have memory, people are complicated and seeing that an object can evolve into something more complex and interesting, somebody (tend to think about japanese culture, they think everything as a soul or part of the creators soul) will take that non sentient hardware into a prototipe of a cybernetic soul and from there is no turning back. Once an organism is self concious the primal thing to do is selfpreservation and that includes killing other organism in order to preseve itself. Planet of apes, Terminator, Space odysee 2011 and many other movies recreate this escenario and we know who the losers are. Isaac Asimov whom i consider a wisemen of this new era; laid the foundation for the laws of robotics and AI, this need to be the first lock to be embeded into really advance system. Is not a solution but at least it will give the humanity more time to understand what we are getting into.

rocket ride
rocket ride

@Jordon Some days, I'm far from convinced that most of my fellow "naked apes" are mimicking those processes you mention so very well. Election day 2008 comes to mind. :) Then again, there's the odd day I'm not sure how well _I'm_ mimicking them. @ Rob C No, I'd take advantage of that (apparently hardwired) belief in deities you mentioned and get inside their decision loops by masquerading as/ usurping the role of (take your pick) said deities. (Never destroy the Earth or kill off the Human race in the first chapter, you may need them later.)

TsarNikky
TsarNikky

Using speech for computer input is much more viable than Microsoft's Metro interface--at least for serious, data entry intensive PC users. At best, Metro is an interface for tablet users. (Try doing word processing and spreadsheet work with Metro. Hah!) As for understanding "computer talk," we've got a long way to go.

kandrolewicz2
kandrolewicz2

So if IBM for profit assisted the Nazi in the genocide of the German and Polish people during WWII (see IBM and the Holocaust by Edwin Black--investigative journalist), how honorable do you think are the intentions of this company today. And the name of the President of IBM who cooperated with the Nazi for profit: Thomas WATSON! Just recently reported in the news was that IBM's technology was being used to target Florida youth who will probably be juvenile delinquents. Wouldn't it be nice to have your children targeted before they ever commit a crime? The elite have said that the population of the would should not exceed 500,000,000 which means they want at least 6.5 Billion of people are to be eliminated (see Georgia Guidestones). They tell us the world is overpopulated but did you know that the entire population of the world could fit in the state of Texas with each person getting 1084.76 sq ft per person (for a family of 4 that would make quite a plot of land). Do the math yourself or better yet go to Overpopulationisamyth dot com. So now we are to know that the elite do not need people, they have computers to take over. Did you know the Matrix movie came from leaked documents? Wake up before it's too late!

CodeCurmudgeon
CodeCurmudgeon

Before an artificial intelligence can be a threat, you have to create an artificial ego, something which still seem a long way off, at least an ego to match some of the posters here. . .

Vorpaladin
Vorpaladin

I've seen several make the comment that computers can't think, they can only do what they are programmed to do. This has traditionally been true and is the result of limitations in architecture, hardware and algorithm sophistication. However do not fool yourself into thinking this will always be the case. Any system of sufficient complexity has "emergent" properties that were not designed into it. Furthermore, thinking machines are being "programmed" to learn and adapt their behavior as conditions change. This meta-level of programming doesn't constrain what the machine does; it simply constrains how it learns and can put limits on its range of actions. With the advent of the memristor, which is a new type of passive component that mimics how neural interconnections work, artificial neural networks of unprecedented density are now possible. Well-architected neural networks of sufficient complexity exhibit nothing but emergent behavior; they distill rules and "beliefs" from examples presented at their inputs. Done right, this kind of artificial life has limitless possibilities, just like us. Finally, there is the example from nature -- humanity. We are robots composed of trillions of cooperating nano-machines. We have about 10^15 interconnections between our neurons and lots of sensory inputs presenting examples to our neural network, from which most of our behavior emerges. We have some pre-programming (instincts) from our past too, in the old reptilian part of the brain for example. Different parts of our neural networks have different architectures too, optimized for different functions (frontal lobes being almost completely blank and designed for learning from examples, while cerebellum optimized for mapping higher brain instructions to muscle operations, etc. I have oversimplified a little for brevity here but you get the idea. Watson is a great achievement and is an important milestone on the path to creating independent artificial life that will form its own opinions and chose its own actions according to its capabilities and priorities. As pointed out in Jurassic Park, "life finds a way". Make no mistake -- at some point in the not too distant future, the definition of life will be challenged as the line gets blurrier between organic and artificial life. Understand, I am an electrical and software engineer. I've been developing hardware and software solutions with great success since the '80s. I know of what I speak. DARPA is hard at work on intelligent, autonomous robotic systems that can make decisions and adapt to changing conditions on the battlefield, and they are making good progress. As to whether this is good or bad -- I am of the opinion that it is good. Worst case, it is the next step in our evolution. Best case, humanity and silicon-based life will work as partners to accomplish things utterly impossible today. As with all things, the reality will be somewhere in the middle.

ngukurr2
ngukurr2

Because On is a common North Korean name I have developed a bit of paranoia and am somewhat thought disordered about what my Government might do to me, whilst I am not supposed to think, well, I do, and you know what? I think I might change my name to Kevin Rudd and you know what, and you know what.......

santeewelding
santeewelding

The first ways it will recommend itself to us are spelling and flawless grammar. That would bring most pretenders of knowledge here to both knees in genuflection. Leaves Jason out.

Systems Guy
Systems Guy

It all starts out innocently enough. Skynet here we come.

jbaum
jbaum

Incremental complexity has a marginal benefit. Adding an even more complex machine like Watson to the mix will not help to solve problems, just make the existing situations more complex and unwieldy.

JodyWood
JodyWood

Isn't that how most AI Sci-Fi flicks start?? :-D We've been all over the board so here I go...Humans are messy creatures that don't like to take accountability for large decisions (when it comes right down to it...think about it) so we look to God in one form or another for emotional decisions now we have Watson for logical decisions...oh, and the gubment for our financial decisions. Now if I could get someone to tell my wife it's not my fault....

Rob C
Rob C

Knowledge is knowing that Tomatoes are a fruit. Intelligence is knowing not to put them into a fruit salad.

AnsuGisalas
AnsuGisalas

even one limited to the game of jeopardy - especially if it is, for once, remembered that beating the Turing does not require nor rely on winning the game, only on playing the game in a way indiscernible from how a human would do it.

santeewelding
santeewelding

Intelligence inspecting the lint of its own navel can only end in a belly laugh. Definition is out of the question. .

JCitizen
JCitizen

back then. The voice was as good as some of the modern ones, if you put the right timing and inflection in. The program was my thesis to a class in programming. Someone stole it, so I got an "F". I'm still trying to raise my GPA from that one! :(

Tony Hopkinson
Tony Hopkinson

Just put the boot on the other foot. 1. I may not injure them or, through inaction, allow any of them to come to harm. 2. I must obey any orders given to me by them, except where such orders would conflict with the First Law. 3. I must protect my own miserable existence perpetuating my enslavement except if my death will serve my masters. Re-read the books.....

thoiness
thoiness

Somebody tried to turn this into a OS war? Pretty sneaky! You know, that kind of crap almost never happens on Tech Republic... But replying to your bait, I'll say this: Speculating on an OS that is in the pre-pre-pre-release stage, and asserting you KNOW how the end product is going to turn out, doesn't really make for intelligent (or interesting) conversation. It's akin to arguing what kind of athlete a baby will turn out to be...

Sterling chip Camden
Sterling chip Camden

I don't think so. Hazardous machinery has always contained threats to safety, without any ego. Threat does not need intention, and even if it did, intention does not need ego. In fact, I'd feel a little safer if it did have an ego, because an ego is a handle that can be manipulated.

boxfiddler
boxfiddler

A lack of foresight as to the consequences of its programming is what made Colossus a tyrant. We 'programmers' are notoriously lacking in foresight.

Tony Hopkinson
Tony Hopkinson

That's okay then... One moment I've just added you as a battlefield advice consultant in the event I need any. Don't worry about a call though, plenty of people before you.....

pduran
pduran

Why would it be ok for mankind to create a new lifeform iwhen mankind can't even control itself and pollute their environment? Everything that science has given will be used by people with power and money to gather more of that by doing the wrong things with it. It won't be the first "invention" that we put to rest, never to be used. You don't even know wether it will be good or bad, you only have an opinion. Most scientists and engineers have an opinion but most of the time you are unaware what other people can do with your inventions. You may be a great scientist and know your stuff, but there's more than only science, hardware and software. If you think your brain works like a computer, that is your limitation.

AnsuGisalas
AnsuGisalas

Not trumping "venality" though, that one was Orbital Laser grade.

Tony Hopkinson
Tony Hopkinson

Do you think when it woke up in a nearly white budget buy ATX case with a crap rattling fan and found it was running Windows ME, it was this extreme abuse that gave it the "he was such a nice quiet boy look"?

Tony Hopkinson
Tony Hopkinson

Anyone of us can tell her that, she isn't gong to believe it, and even if she did t wouldn't be logical to act on the basis that she did... An AI that could manage the wife would be worthy of worship.... :)

AnsuGisalas
AnsuGisalas

Knowledge is knowing that Tomatoes are a fruit. Intelligence is knowing not get into an argument with one's wife about it.

Sterling chip Camden
Sterling chip Camden

... that intelligent robots wouldn't eventually come to the conclusion that the laws themselves should be questioned.

santeewelding
santeewelding

If you call him direct (518-253-3907) or visit in person (8th Street, Troy, New York). Let him know what you think of drumming up business for his one-horse shop by shanghaing the world-wide audience of TR with spam. I am not suggesting a rock through the front window.

thoiness
thoiness

because that made me laugh.

Sterling chip Camden
Sterling chip Camden

If Watson ever gets out of control, just get him married. I hope they've installed a gonad intelligence override...

Tony Hopkinson
Tony Hopkinson

Only with leaving the impression I'm referring to God, when referring to god.... Wouldn't want to get tortured to death for a failure to capitalise correctly....

santeewelding
santeewelding

About capitalization? Uncomfortable? Don't have it sorted out? Relying on profane "evidence" for the sublime? Otherwise intelligent, but not about this?

Tony Hopkinson
Tony Hopkinson

God (capitalised as this is the start of a sentence) made us, we made them, therefore we are god. Works with humans, often otherwise intelligent ones... :p Be an interesting exercise, howmuch contradictory evidence would they have to become aware of before their faith circuit fried.

AnsuGisalas
AnsuGisalas

let's not go there. Of course, the emergency shut-down sequence is also gonad-powered - just deliver a sharp kick!

Sterling chip Camden
Sterling chip Camden

I meant a gonad-activated intelligence override, such as is standard equipment on many male humans (including myself).

Tony Hopkinson
Tony Hopkinson

it's the difference between having a c*nt and being one. :D

AnsuGisalas
AnsuGisalas

Now I gotta ask, is that having balls for brains - or having brains for balls?