IBM Watson CTO on Why Augmented Intelligence Beats AI

This episode of Fast Forrad was recorded in the IBM Watson Experience Center here in New York Urban center. My guest was Rob High, the Vice President and Master Engineering Officer of IBM Watson.

Fast Forward Bug Art High works across multiple teams within IBM, including engineering, development, and strategy. He is one of the most lucid thinkers in the space of artificial intelligence, and our conversation covered many of the style that applied science is reshaping our jobs, our society and our lives. Read and picket our conversation below.

Dan Costa: What is the dominant misconception that people accept about artificial intelligence?

Rob Loftier: I recollect the almost mutual problem that we're running into with people talking nearly AI is they still live in the world where I call back Hollywood has amplified this idea that cognitive computing, AI, is virtually replicating the human being mind, and information technology'southward really not. Things like the Turing test tend to reinforce that what we're measuring is the thought of AI being able to compete with fooling people into believing that what y'all're dealing with is another human, just that'southward actually not been where we have establish the greatest utility.

This even goes back to, if y'all expect at nearly every other tool that has ever been created, our tools tend to exist most valuable when they're amplifying the states, when they're extending our reach, when they're increasing our strength, when they're allowing usa to do things that we can't do by ourselves as human being beings. That's actually the way that we need to be thinking near AI besides, and to the extent that we actually call it augmented intelligence, not artificial intelligence.

Let's talk a fiddling bit most that shift, because information technology's an entirely new type of computing. It'due south the development of calculating from what we both grew up with, a programmatic computing where y'all would utilise ciphering to reach and answer using a very complex process, to cognitive computing, which operates a little differently. Can you explicate that transition?

Probably the biggest notable departure is that it's very probabilistic, whereas programmed computing is really about laying out all the conditional statements that define the things that you're paying attention to and how to respond to them. Information technology's highly deterministic. It's highly mathematically precise. With a classic programmed figurer, you can design a piece of software. Because you lot know what the mathematical model is that it represents, you can test information technology mathematically. You tin prove its correctness.

Cerebral computing is much more probabilistic. It'due south largely about testing the signals of the spaces that we're focused on, whether that is vision or speech or language, and trying to observe the patterns of meaning in those signals. Even then, there's never absolute certainty. Now, this is in function because that'due south the mode it's computed, but also because that's the nature of human being experience. If you think almost everything that we say or see or hear, taste or affect or smell or anything that is part of our senses, we as homo beings are always attempting to evaluate what that actually is, and sometimes we don't get it right.

What'southward the probability that when I heard that sequence of sounds, it really meant this word? What's the probability that when I saw this sequence of words it meant this statement? What'south the probability that when I see this shape and an image that I'thou looking at that information technology is that object? Even for man beings, that's a probabilistic trouble, and to that extent it's always the way that these cerebral systems work every bit well.

If somebody comes to you and they have a problem that they want to solve, they think that in that location is a cognitive computing solution to that, they come to Watson, they say, "Look, nosotros're going to use Watson to try and solve this trouble." Out of the box, Watson doesn't do very much. They need to teach information technology how to solve their problem. Can you lot talk about that onboarding process?

Actually, we should talk about two dimensions of this. Ane is that some time agone we realized that this thing chosen cerebral computing was really bigger than u.s.a., it was bigger than IBM, it was bigger than any one vendor in the industry, it was bigger than any of the 1 or two different solution areas that we were going to exist focused on, and we had to open it up, which is when nosotros shifted from focusing on solutions to really dealing with more of a platform of services, where each service really is individually focused on a different part of the problem space. It'due south a component that, in the case of speech, is focused strictly on the problem of trying to take your spoken language and recognize what words you lot've expressed in that spoken communication, or accept an image and endeavor and identify what's in the epitome, or take language and attempt to understand what its pregnant is, or accept a chat and participate in that.

Watson

First of all, what nosotros're talking about at present are a set of services, each of which do something very specific, each of which are trying to deal with a different function of our man experience, and with the idea that anybody building an application, everyone that wants to solve a social or consumer or concern problem can do that past taking our services, then composing that into an application. That's point ane.

Betoken two is the 1 that yous started with, which is, all right, at present that I've got the service, how practise we get it to practice the things nosotros want it to do well? The technique really is one of teaching. The probabilistic nature of these systems is founded on the fact that they are based on machine learning or deep learning, and those algorithms have to be taught how to recognize the patterns that stand for pregnant within a set of signals, which you practise past providing data, data that represents examples of that situation that you've had earlier where you've been able to characterization that as saying, "When I hear that combination of sounds, it ways this word. When I see this combination of pixels, it means that object." When I had those examples, I can now bring you to the cognitive system, to these cognitive services, and teach them how to do a better job of recognizing whatever information technology is that nosotros want information technology to do.

I recollect ane of the examples that illustrates this actually well is in the medical space, where Watson is helping doctors brand decisions and parsing large quantities of information, just and so ultimately working with them on a diagnosis in partnership. Can you talk a footling bit about how that grooming takes identify and so how the solution winds upward delivering meliorate outcomes?

The work that we've done in oncology is a expert example of where actually it's a composition of multiple different kinds of algorithms that, beyond the spectrum of work that needs to exist performed, are used in different ways. We commencement with, for example, looking at the medical record, looking at your medical record and using the cognitive organization to look over all the notes that the clinicians have taken over the years that they've been working with you and finding what we call pertinent clinical information. What is the information in those medical notes that are at present relevant to the consultation that yous're virtually to go into? Taking that, doing population similarity analytics, trying to notice the other patients, the other cohorts that take a lot of similarity to yous, because that's going to inform the doctor on how to call up well-nigh different treatments and how those treatments might be appropriate for you and how yous're going to react to those treatments.

And so we become into what we telephone call the standard of care practices, which are relatively well-defined techniques that doctors share on how they're going to treat different patients for different kinds of diseases, recognizing that those are really designed for the average person. Then nosotros lay on top of that what we phone call clinical expertise. Having been taught by the best doctors in unlike diseases what to look for and where the outliers are and how to reason virtually the different standard of care practices, which of those is about appropriate or how to take the different pathways through those dissimilar intendance practices and now apply them in the all-time way possible, merely finally going in and looking at the clinical literature, all the hundreds of thousands, 600,000 articles in PubMed about the advances in science that have occurred in that field that are relevant to now making this treatment recommendation.

All those are different aspects of algorithms that we're applying at different phases of that process, all of which accept been taught by putting some of the best doctors in the globe in front of these systems and having them use the system and correct the system when they see something going wrong, and having the system larn essentially through that use on how to improve its ain performance. We're using that specifically in the case of oncology to help inform doctors in the field about handling options that they may not be familiar with, or even if they accept some familiarity with it may not have had any real experience with and don't really understand how their patients are going to respond to it and how to get the nearly effective response from their patients.

What that basically has washed is democratized the expertise. We can take the all-time doctors at Memorial Sloan Kettering who had the do good of seeing literally thousands of patients a year around the aforementioned disease from which they've developed this tremendous expertise, capture that in the cognitive organisation, bring that out to a community or regional dispensary setting where those doctors may not have had as much time working with the same disease across a large number of dissimilar patients, giving them the opportunity to benefit from that expertise that's now been captured in the cerebral system.

IBM Watson

I think that idea of distributing that expertise, outset of all, capturing it is a non-trivial job, but and so once you've washed that, being able to distribute it actually beyond the planet, you're going to accept the expertise of the best doctors at Memorial Sloan Kettering beingness able to exist delivered in China, in India, in small clinics, and I recall that's pretty extraordinary.

It has a tremendous social bear upon on our welfare, on our health, on the things that volition benefit united states of america as a society.

On the flip side, the thing that concerns people about artificial intelligence is that information technology's going to replace people, it's going to replace jobs. It's tied into the automation movement. The thing that strikes me is, staying in the medical space, radiologists. Radiologists look at hundreds and hundreds of slides a day. Watson or an AI-based organisation could replicate that same type of diagnosis and epitome assay. X years from now, practice you think at that place are going to exist more or fewer human radiologists employed in the Usa? What's the affect on industries like that?

The impact is really nigh helping people do a better job. It's really virtually ... have it in the case of the dr.. If the doc can now make decisions that are more than informed, that are based on existent evidence, that are supported by the latest facts in science, that are more tailored and specific to the individual patient, information technology allows them to actually do their job ameliorate. For radiologists, information technology may let them to see things in the image that they might otherwise miss or get overwhelmed by. It's not near replacing them. Information technology's about helping them practice their job better.

It does have some of the same dynamic that every tool that we've always created in society. I similar to say if you become dorsum and look at the last x,000 years of modernistic guild since the advent of the agricultural revolution, we've been as a human society building tools, hammers, shovels, hydraulics, pulleys, levers, and a lot of these tools take been virtually durable when what they're really doing is amplifying human being beings, amplifying our force, amplifying our thinking, amplifying our achieve.

That's really the way to think about this stuff, is that it volition have its greatest utility when it is allowing united states to practise what nosotros exercise ameliorate than we could by ourselves, when the combination of the human and the tool together are greater than either 1 of them would've been by theirselves. That's actually the way nosotros think about it. That'southward how nosotros're evolving the technology. That's where the economical utility is going to exist.

I completely concord, only I practise think there'south going to be industries that are obviated because of the efficiency introduced by these intelligent systems.

They're going to exist transitioned. Yep, they're going to be transitioned. I don't desire to diminish that signal by saying it this way, but I also want to be sure that we aren't thinking near this as the elimination of jobs. This is about transforming the jobs that people perform. I'll give you an example. A lot of word about how this may take away jobs in the telephone call center. Well, estimate what? There'due south a lot of work that call eye agents practise that they don't demand to be doing, they don't like doing, that takes away from their ability to practise things that are more than interesting.

The churn that nosotros encounter in call centers is largely driven by the fact that if you retrieve virtually the job of being a call center agent, you're sitting at the end of telephone call listening to irate customers all solar day long request the same question over and over once again, and it'southward hard to get home at nighttime feeling really good nigh what you lot did that day. It'southward hard to brag to your friends and family about this job that you lot have and how proficient you are at doing information technology when that's the state of affairs yous're in.

If we tin can become the cognitive organisation through a conversational agent to offload some percentage, let'due south say 30 pct of those calls coming in, and answering the customers' about common and pressing questions rapidly, efficiently, and take care of that mundane work, then what'south left after all that'southward been taken care of are the kinds of questions that people have that inherently require more of a human being affect that and then yous're going to plow over to that phone call eye agent. The problem that they're dealing with for that customer is more than interesting, more challenging, requires them to have more than intellectual endeavor put into information technology, but as well they're dealing with a customer who's been satisfied. They're coming in a trivial bit happier. They're non coming in all irate most their trouble.

For the call center agent, it actually has improve their job. Information technology really makes information technology possible for them to do their job improve and be more fulfilled by that. In the meantime, for the customer, for the consumer, they got their most pressing issues solved rapidly. They're not sitting on agree for 10 minutes. They're not waiting for the get routed to the right person with just the right knowledge. They're getting the information they demand most readily and able to move on with their life with probably a amend decision, certainly ameliorate information or at least more consistent information. Information technology actually benefits both sides of that equation.

It'south interesting. Some of the demos I saw today is that the call heart applications can conceptualize and discover the emotional state of the people that are calling in pretty effectively, then it's not but transactional. It can really read the state of the person on the other end of the line pretty well.

Which is actually essential if y'all think well-nigh [it]; a conversation has two elements to it. Ane is that what people say to begin with is by and large not what they're really there for. If I say, "What's my balance?" well, that's not really my problem. Yeah, I need to know my account balance, I need to know how much money I have, but my problem is I'thousand trying to buy something, or I'thou trying to figure out how to get money in the right position to pay my bills this month, or I'g trying to save up for my kids' teaching. My problem is bigger than that first question I asked, and a conversation should be nearly getting to that real problem.

The second common characteristic of a chat is that typically it carries a sort of emotional arc to information technology. People come in in a sure emotional country, and office of the chat is to motion them through an emotional shift that oftentimes means moving them from existence angry to now being satisfied. In some conversations, we might get into it. Information technology might actually become a little heated. You see an emotional arc that starts out maybe at-home and then moves to a more contentious discussion that eventually and then gets resolved.

Being sensitive and aware of emotional state in the parties involved is an important part of being effective in that conversation.

What are some of the other applications that you think are actually transformative that are available today?

I think that [with] whatever of them, what nosotros're doing is engaging the user, the customer, in a way that results in inspiring them. For me, ultimately, and once more going back to conversations as an example, typically when human beings become into a conversation, we come up to the table with an idea. You lot take an thought. I have an thought. That starting thought is the first of the conversation, and over the form of the conversation nosotros evolve those ideas. We blend them. We merge them. Nosotros maybe discount them or amplify them. We evolve to a point where coming out of the conversation we have a better idea, hopefully. Ideally.

To do that, there has to exist non merely the give and take, but an element of how do you inspire somebody? How do you cause people to actuate their imagination? How exercise you cause them to remember about something they hadn't idea well-nigh before or meet something in a calorie-free they hadn't thought of before or to encounter another betoken of view that takes them down a path that they didn't even know to think well-nigh, to ask questions they're not thinking to inquire? Those are the examples, those are the situations that I recollect are most promising and will accept the greatest do good for people.

Is that happening today, or is that something that needs to happen downward the line as the technology evolves?

No, it's happening. We have examples of that happening now. In fact, going back to oncology as an exemplar, for the best doctors in the world, the handling options that are being presented may be obvious to them for the virtually office. There may be one out of ten cases where they might say, "Well, wait a minute, that was an interesting idea." It won't exist as often, simply, like you lot said before, if we take that now out to community settings, regional settings, and in areas where there aren't that levels of expertise, the fact that the system can introduce new ideas, new treatment options, information technology's really about introducing new ideas. We're seeing that already.

Then, of course, moving beyond what I recall has get the archetype chatbot scenario that I remember some of us are outset to see in different examples to now a state of affairs where if somebody gives a credit carte du jour fraud alert on their credit card and they get to a chatbot today, it might be simply but, "Was that transaction something that you did or not? If it is, then fine. If not, then nosotros're going to practice something about canceling the transaction," into now, "Okay, y'all need a new credit card. Where's the all-time place to get it to you? Should we mail service it you? Should we not mail information technology to you? Oh, you're getting ready to continue this trip. Then clearly we're non going to be able to mail it to you. We've got to get it to you faster than that.

"Oh, you're going overseas. Mayhap there's a credit carte du jour choice here that you weren't exposed to before, didn't know about, where we handle currency exchanges in your favor ameliorate. Oh, you're using this for business. This is an overseas trip. You lot're using this for business concern expenses. Well, here's a credit menu that has an interest rate that'south more appropriate for that." These are all very unproblematic examples, but each of them are opening up a new set of ideas that doesn't typically happen in your simple chatbot today and yet can actually be very empowering for human beings.

Watson VA Cancer

The interesting indicate in that location is that as you're going through all of those options, in the past that would be a script. There would exist a script with a couple of branches. It would exist predefined in advance. It'southward a very different thing when a chatbot does information technology that'southward actually reacting to the information you give and the information you've already given and leading y'all down paths that accept non been scripted. It knows that yous're traveling, but y'all haven't necessarily told it. It found that information from your e-mail history.

It tin find things almost you information technology discovered along the way.

Nosotros talked about oncology because it's a great example. We talked about chatbots because virtually people have had some interaction with them. But this is a applied science that really scales across every industry. It's difficult to remember of an industry that won't take some kind of cerebral component to it. Are there any examples that are simply style out there that people haven't thought virtually still?

The thing that's astonishing to me is how every single day somebody'southward coming upwards with another new idea. That's why I recollect we're in such a very interesting stage, because by having focused on decomposing what we have in terms of cognitive capabilities into building cake services, it'southward actually freeing up people to utilise their imagination and go pursue ideas that we've never really considered before, whether that is using visual recognition to survey the landscape.

In California, for case, a company there is using visual recognition to look at the topography and the topology and recognize in the image the departure betwixt a concrete surface, an asphalt roof surface, a grass surface, copse and shrubs and these things, to estimate how much water is existence consumed and where there may exist water leaks and things that could be done to better the efficient use of water, equally an instance.

Or, in the legal arena, using these things to go off and help lawyers read through with literally millions and millions of pages of background material that is similar finding the needle in a haystack. Where's that one piece of paper that is really relevant to this particular instance? Trying to sort through all that. The opportunities are just enormous.

I recall that ane of those qualifications is having big quantities of data that need to exist parsed through. You lot talked virtually medical records and being able to scan the medical records for the relevant information. Those records over the course of your lifetime could be many of hundreds of pages long. That'due south the thing that, maybe your family physician has an inkling of that, but they're not going to remember all of information technology, whereas the system never forgets.

Yeah. A doctor may have 5, maybe 10 minutes to await through that medical history before coming and consulting with you, and yet at that place'south all kinds of very relevant information that may be in your history, your past, that under any other circumstances they would miss just considering they don't have the time, that if they had that would make a difference.

Call back about a state of affairs where if a adult female had told her doctor that her mother but died of chest cancer ii years ago. Well, chances are that doctor will have noted that in that record, but at this moment, if this woman's coming in presenting a lump in her chest, and if that doctor doesn't see that, well, that's a very important piece of missing data. Now, maybe they'll rediscover that by talking to the patient, but maybe non. Exercise yous actually want to take the risk of not having known that when something like that is so germane?

The overarching feature for where this stuff tends to exist useful is you mentioned where there'southward lots and lots of data. Aye, but really it'southward when any of those aspects of who nosotros are as human beings, where what our cognitive adequacy begins to attain its limit. We're practiced at reading. We can read something. We tin assimilate information technology. We can adapt to the information and make utilise of that in very powerful ways equally human beings. But nosotros're not very skillful at reading lots of data. We tin't read more than ... The idea of reading tens of thousands, hundred thousand, millions of pages of literature in a day is and so far beyond our capacity.

The question becomes, as we grow into a world where the amount of information that'south produced on a daily ground is growing exponentially, how much more than of that information are we not making use of that has information in it, has that footling tidbit of information that'due south admittedly critical to the decision we need to make are we non getting to? If it'south not the corporeality of information we read, it'due south: How much practise we assimilate? How much are nosotros able to call up? Are nosotros able to run across the fiddling patterns that are relevant in that information to our decisions?

In that location are lots of things that we as human beings are good at. There's also a lot of things that we're not very proficient, and that's I think where cognitive computing really starts to make a huge difference, is when it's able to span that altitude to make up that gap.

It seems pretty articulate this is the world that we're moving into. How prepared are we? What practice you lot look at our education system, our economic system, our political structures? How well prepared are we to live in a world with this type of cognitive computing equally a component?

It's interesting. This draws on one of the key value points that we possess as human beings, which is our power to adapt. If you wait at it in purely detached terms, where is this going, and if we were to bound forwards ten years and expect at it and say, "Where will nosotros be 10 years? Are we prepared for that?" the answer is going to probably be, no. There'southward a lot more than that we take to do. Only human beings take this remarkable power to adapt on the fly and grow with the changes that are occurring around them.

Think back 10 years ago when the smartphone was really just starting to become available to us, let lone popular, and how much change nosotros take gone through every bit a order over the last x years. Think about what your life is like on a daily basis with and without your smartphone. We tin can mutter about how much it may exist taking away from other experiences, and that may be truthful, but the indicate is, nosotros didn't spend a lot of time x years ago fretting over, were we prepared as a society, fifty-fifty though in fact we've gone through a lot of changes over the last 10 years that we probably weren't fully aware of as we assimilated this modify in applied science and started making utilize of it in very constructive ways.

There'south a lot that we take to practice. There's a lot that we're going to be doing over time, a lot of growth that we'll become through, a lot of education and politics and other things that we have to go through changes on, simply we will.

We'll get to my last questions. What technological trend concerns you the most? Is there annihilation that keeps y'all up at dark?

I recollect that the biggest concern I take right at present is people do need to take responsibility. We as engineers and providers of engineering, consumers of engineering science, people who have responsibleness for regulating technology, really do need to be witting and think through now what we want to do to protect ourselves and prepare ourselves for the changes that are occurring. It won't exist considering we won't adapt to information technology. Nosotros will. The problem is of course, in the process of adapting it, we also won't be conscious of what that is doing and how that's affecting the states and where people may be exploiting that technology in ways that we don't prefer, that nosotros aren't comfortable with, or in retrospect nosotros won't necessarily want.

I exercise think that we need to be conscious and thinking about what we exercise and we don't want to have happen in our lives with this technology. Specifically, vendors in particular, we as the suppliers of this engineering, and the people who are consuming these technology components and building applications out of it should at this moment presume responsibility for our ethical behavior or behaviors that are born from ethical values.

As an example, we strongly recommend to whatever of our application developers, whatever of the institutions that are creating applications using these technologies, that they be very transparent with their end users nigh the fact that this is a cognitive awarding, it's a computer, and non endeavour to masquerade every bit a real man existence, for example. Don't pretend. Don't let this matter pretend.

Don't imitate.

Don't imitate information technology and don't let your customers ever be mislead into believing that this matter is a existent person. Ethically, information technology's wrong. I retrieve it creates the risk of vulnerability. A human being who is interacting with a human being being tin can make certain assumptions about our flaws, about our inability to really retain a lot of information, where when dealing with a cognitive system, we need to be mindful that the people who are providing that cognitive solution take a responsibleness to the privacy and protection of the information that nosotros supply information technology. We shouldn't be ever forgetful of that fact.

In terms of engineering on the upside, what technology do yous use every day that only inspires wonder? What'south changed your life?

I think the fact that I can now get access to information that, even if I could get information technology in the internet, we've had information available to us in the internet for a long fourth dimension, just oftentimes nosotros stop trying to get that information because it's overwhelming. I was out looking at some camera equipment, and just trying to make decisions about the merchandise-offs between unlike cameras-

I'll send you a link to our buyers guide.

There y'all go. It gets overwhelming, and yet you take to rely on other people to provide that communication for you and assume that they've done the enquiry for yous, only even then, they're doing so based on some assumptions they've made about what you demand and what yous care about. At some indicate yous merely just give up and you say, "Okay, fine, just tell me what to do, I'll do it." Or you go to a whole agglomeration of websites and yous see all these opinions and it just gets disruptive and contradictory and then you say, "Well, the heck with all them. I'm just going to become with what feels good to me."

Now, because these systems tin accumulate and assimilate and organize vast quantities of information, even for the people who are making recommendations, even for the advisors, it benefits them because it helps them do a better job. A fashion I similar to say it is it doesn't practise our thinking for us, it does our research for united states of america so nosotros can do our thinking ameliorate, and that'southward true of u.s.a. equally finish users and information technology'south truthful of advisors. It'south truthful of anybody who'due south in that office of being an annotator.

I think of the application, considering nosotros're e'er trying to help people make buying decisions. We're not far from a organisation that could await at all the photos that you lot've taken over the last five years, see that you like to do wildlife photography or closeups of flowers, and then make a camera recommendation based on the pictures that you have.

That'southward right. Flamingos. I don't know why.

This is the best camera for taking pictures of flamingos.

Flamingos, right.

We're almost there. The technology exists, it just hasn't been programmed yet.

Yeah.

Or taught, every bit we do these days. Rob High, thanks so much for doing this.

Thanks very much.

For more Fast Forrard with Dan Costa, subscribe to the podcast. On iOS, download Apple's Podcasts app, search for "Fast Forward" and subscribe. On Android, download the Stitcher Radio for Podcasts app via Google Play.