WEBVTT - AI at IBM Think 2019 0:00:04.120 --> 0:00:07.480 Get in text with technology with tech Stuff from stuff 0:00:07.520 --> 0:00:13.600 works dot com. Hey there, and welcome to tech Stuff. 0:00:13.600 --> 0:00:16.200 I'm your host job in Strickland. I'm an executive producer 0:00:16.239 --> 0:00:17.880 with How Stuff Works in my Heart Radio and I 0:00:17.960 --> 0:00:22.200 love all things tech. And if today's episode sounds a 0:00:22.200 --> 0:00:25.360 little different one, it's a special episode two. I'm recording 0:00:25.360 --> 0:00:29.560 it on location in San Francisco, California, so you might 0:00:29.640 --> 0:00:34.040 occasionally hear some traffic noises, some hotel noises. Maybe you'll 0:00:34.080 --> 0:00:37.400 hear a bell of the famous San Francisco trolley. Perhaps 0:00:37.400 --> 0:00:41.480 you'll even hear bagpipers, because we did. But I'm here 0:00:41.680 --> 0:00:45.360 in San Francisco for a specific reason. IBM invited me 0:00:45.479 --> 0:00:48.240 to fly out here and attend the Think two thousand 0:00:48.280 --> 0:00:51.400 nineteen conference and really get an up close and personal 0:00:51.479 --> 0:00:54.160 view of some of the innovations and services the company 0:00:54.200 --> 0:00:58.160 is rolling out all to their clients, their business partners, 0:00:58.160 --> 0:01:00.760 and I really wanted to share with you my own 0:01:00.840 --> 0:01:05.520 takeaways from this event. Now, before I jump into all this, 0:01:06.160 --> 0:01:09.600 IBM is a business to business entity, meaning that if 0:01:09.640 --> 0:01:13.960 you're an average Jonathan like me, you rarely deal directly 0:01:14.319 --> 0:01:17.679 with IBM. But the company is one of those leading 0:01:17.800 --> 0:01:21.880 entities that provides the tech that other companies use in 0:01:21.959 --> 0:01:25.640 order to do their business. So while it may or 0:01:25.680 --> 0:01:27.880 may not be obvious, there's a lot of stuff that 0:01:27.920 --> 0:01:30.399 we encounter in our day to day lives that's powered 0:01:30.520 --> 0:01:34.640 by IBM. This episode is the first of four special 0:01:34.680 --> 0:01:38.280 episodes about the conference and the technology and innovations that 0:01:38.319 --> 0:01:42.400 are at the bleeding edge of deployment, and today we're 0:01:42.440 --> 0:01:47.000 going to focus on artificial intelligence, something that you could 0:01:47.360 --> 0:01:51.120 argue is almost synonymous with IBM. So you guys know 0:01:51.440 --> 0:01:54.360 that AI is one of my favorite topics to talk about, 0:01:54.440 --> 0:01:56.279 and it can be easy to fall into the trap 0:01:56.360 --> 0:02:00.600 of thinking about AI as some sort of nebulous intelligence 0:02:00.640 --> 0:02:04.120 living in a machine, But when you strip away the 0:02:04.240 --> 0:02:07.360 veil of mystery, you'll see that AI is just another 0:02:07.400 --> 0:02:11.160 part of computer science. It might rely on one architecture 0:02:11.320 --> 0:02:14.880 over another, or it might require an artificial neural network approach, 0:02:14.960 --> 0:02:17.839 depending upon the application, but really it just comes down 0:02:17.880 --> 0:02:21.480 to a series of special algorithms designed to handle information 0:02:21.800 --> 0:02:25.280 in a way to allow a computer to make decisions. 0:02:25.720 --> 0:02:30.520 It's sophisticated and it's fascinating, but it's not magic. However, 0:02:30.919 --> 0:02:34.320 you might be forgiven for thinking of it as magic 0:02:34.480 --> 0:02:37.679 if you happen to witness the exchange between IBM s 0:02:37.720 --> 0:02:44.120 AI System Project Debater and Grand Champion debater Harish Natarajan. 0:02:44.639 --> 0:02:47.519 The debate between man and machine happened a day before 0:02:47.560 --> 0:02:51.480 the official start of the conference. The two participants of 0:02:51.520 --> 0:02:55.560 the debate we're not told about their topic until fifteen 0:02:55.600 --> 0:02:59.280 minutes before the debate was to begin, and then they 0:02:59.280 --> 0:03:02.639 were given their stance on what that topic was. They 0:03:02.680 --> 0:03:07.519 each had four minutes to establish their positions on the subject. 0:03:07.919 --> 0:03:10.799 Then after a short break, they had another four minutes 0:03:10.840 --> 0:03:14.639 to offer rebuttal of their opponent's stance, and then one 0:03:14.720 --> 0:03:18.640 short break later they had two minutes to summarize their arguments. 0:03:19.080 --> 0:03:23.280 The debate topic turned out to be preschools should be subsidized. 0:03:23.720 --> 0:03:27.440 Project Debater, who, by the way, has a gender. Project 0:03:27.520 --> 0:03:31.600 Debater is a she. She was given the pro stance 0:03:31.880 --> 0:03:35.960 on that particular argument and Mr Natarajan got the counter 0:03:36.160 --> 0:03:41.320 stance the the the idea that preschools should not be subsidized. 0:03:41.880 --> 0:03:44.520 I am not going to go through a blow by 0:03:44.600 --> 0:03:48.000 blow of the debate. For one thing, you can actually 0:03:48.040 --> 0:03:51.480 listen to it yourself. Intelligence Squared, which is a show 0:03:51.560 --> 0:03:55.160 dedicated to civil debate on a wide array of topics 0:03:55.520 --> 0:03:59.760 played host to this particular special debate. I urge you 0:03:59.800 --> 0:04:02.960 to seek out that podcast or a video of the 0:04:03.000 --> 0:04:06.160 debate if you want to see how it unfolded for yourself, 0:04:06.720 --> 0:04:09.080 bit by bit. I really just want to talk more 0:04:09.080 --> 0:04:14.040 about the process that was involved. So to debate, no 0:04:14.080 --> 0:04:17.359 matter what you are, whether you're human or machine, you 0:04:17.400 --> 0:04:20.400 need to have an understanding of what it is you're 0:04:20.440 --> 0:04:25.560 either arguing for or against, which is a pretty obvious statement, 0:04:25.560 --> 0:04:27.760 but I feel like I have to lay it out 0:04:27.839 --> 0:04:30.839 that way. You need to be able to form an argument, 0:04:31.360 --> 0:04:34.360 and you have to be able to support that argument logically. 0:04:34.640 --> 0:04:37.559 You want to build your argument so that one part 0:04:37.720 --> 0:04:41.520 leads inevitably into the next part and it all supports 0:04:41.560 --> 0:04:44.440 the stance you have, whether it be for or against 0:04:44.520 --> 0:04:48.200 a particular proposal. This is a non trivial task for 0:04:48.240 --> 0:04:52.640 a human being, and it is an incredible challenge for computers. 0:04:53.160 --> 0:04:57.479 Project debater Or has about ten billion sentences worth of 0:04:57.600 --> 0:05:00.760 data stored in its memory, So when a gets a topic, 0:05:01.200 --> 0:05:03.880 first it has to scour all of the information that 0:05:04.120 --> 0:05:07.240 is in its memory banks and look for relevant information 0:05:07.480 --> 0:05:09.960 related to that topic. Then it has to go a 0:05:10.000 --> 0:05:13.760 step further. It can't just pull up any random information 0:05:13.920 --> 0:05:16.640 about the topic. It has to understand that the information 0:05:16.640 --> 0:05:20.320 actually supports its argument. That is, the computer has to 0:05:20.360 --> 0:05:24.359 make sure it is picking information that is aligned with 0:05:24.440 --> 0:05:29.160 its debate position and not actually against its debate position. 0:05:29.720 --> 0:05:33.320 This falls into the field of natural language processing, and 0:05:33.320 --> 0:05:36.040 I've talked a lot about this too, but in short, 0:05:36.480 --> 0:05:39.320 this describes the area of computer science in which we 0:05:39.400 --> 0:05:42.400 try to find ways for machines to suss out the 0:05:42.520 --> 0:05:47.560 meaning from actual human language. At the basic level, computers 0:05:47.720 --> 0:05:52.080 communicate in machine code and we communicate in human languages. 0:05:52.480 --> 0:05:56.960 Machines don't natively understand human language, just as machine code 0:05:56.960 --> 0:06:00.240 would appear to be nonsense to us. The journey to 0:06:00.320 --> 0:06:04.800 creating powerful systems that use natural language processing to figure 0:06:04.800 --> 0:06:08.640 out the meaning of words, whether they're written or they're spoken, 0:06:09.000 --> 0:06:11.479 it's been a really long one, and many people have 0:06:11.600 --> 0:06:16.000 made advancements, sometimes from completely different perspectives. We've got a 0:06:16.000 --> 0:06:18.800 lot better at this in general, but it's still a 0:06:18.880 --> 0:06:22.160 challenging problem. So think about Google Search for a second. 0:06:22.680 --> 0:06:25.640 When you search for a topic, you type your search 0:06:25.760 --> 0:06:28.120 terms into Google, and then you look at the results. 0:06:28.440 --> 0:06:32.200 You typically get a pretty wide variety of responses. Some 0:06:32.360 --> 0:06:34.760 of them are going to be more relevant than others. 0:06:35.040 --> 0:06:37.479 You might even get a few that aren't relevant at all. 0:06:37.920 --> 0:06:41.680 Google's algorithms attempt to guess at which responses will be 0:06:41.720 --> 0:06:45.120 the most relevant based on your search and sometimes on 0:06:45.240 --> 0:06:49.760 some supplemental information like your search history. But sometimes the 0:06:49.839 --> 0:06:52.400 results aren't ordered in a way that you would prefer. 0:06:52.520 --> 0:06:55.240 You might get an okay response at the top and 0:06:55.400 --> 0:06:58.240 maybe a better one or more relevant one two or 0:06:58.279 --> 0:07:01.560 three spots down. Typically it ends up being on the 0:07:01.600 --> 0:07:04.039 first page, but sometimes it can even be buried lower 0:07:04.120 --> 0:07:07.680 down in the search results. Project debater can't just do 0:07:07.839 --> 0:07:11.400 a simple search and return on key terms, or else 0:07:11.480 --> 0:07:15.000 it might end up spouting out gibberish. It could string 0:07:15.080 --> 0:07:18.200 together two or three sentences that contradict each other that 0:07:18.200 --> 0:07:21.160 wouldn't do anyone any good. And that leads me to 0:07:21.160 --> 0:07:24.400 another point. It's not just good enough to grab relevant 0:07:24.440 --> 0:07:27.640 information that aligns with the argument stance. That's necessary, but 0:07:27.680 --> 0:07:31.120 it's not enough. Those statements have to be ordered properly. 0:07:31.160 --> 0:07:34.440 You have to build support for your stance. You have 0:07:34.480 --> 0:07:38.000 to have this logical progression. A good argument needs that 0:07:38.320 --> 0:07:40.800 from the opening to the closing, so you need to 0:07:40.840 --> 0:07:44.200 make sure there's a flow of information. Otherwise all you'll 0:07:44.200 --> 0:07:46.800 get is a series of relevant points, but they're in 0:07:46.800 --> 0:07:50.119 no particular order, and you have no transitions from point 0:07:50.120 --> 0:07:52.760 to point. It would be jarring and it would be ineffective. 0:07:53.480 --> 0:07:58.400 Project Debater could also support arguments with evidence, which is 0:07:58.520 --> 0:08:01.400 kind of cool. Throughout the debate, we heard the system 0:08:01.520 --> 0:08:05.160 site various studies and quote experts in the field to 0:08:05.240 --> 0:08:10.360 provide support for her stance. This was pretty compelling stuff, 0:08:10.600 --> 0:08:14.160 and this is where Project Debater could be incredibly helpful 0:08:14.200 --> 0:08:18.920 for people who want to argue for or against well anything. Really, 0:08:19.520 --> 0:08:22.320 it's the one area I would say that Project Debater 0:08:22.640 --> 0:08:27.880 had an enormous advantage over the human champion. Harrish. Natarajan 0:08:28.160 --> 0:08:32.000 understands how to create a logical, persuasive argument and how 0:08:32.040 --> 0:08:35.560 to find weaknesses in the arguments of opponents, but he 0:08:35.600 --> 0:08:39.400 can't research a library's worth of information in fifteen minutes 0:08:39.440 --> 0:08:43.760 in preparation for a debate, but Project Debater can. However, 0:08:43.760 --> 0:08:47.000 I wouldn't feel too badly for Harrish. He held the 0:08:47.040 --> 0:08:50.840 advantage in lots of other ways. For example, during the debate, 0:08:51.040 --> 0:08:54.280 he brought up a criticism of Project Debater's argument, pointing 0:08:54.280 --> 0:08:57.720 out that one of her conclusions was based without first 0:08:57.880 --> 0:09:00.880 establishing the evidence needed to support were in it. In 0:09:00.920 --> 0:09:04.160 a debate between human champions, you would likely hear during 0:09:04.160 --> 0:09:07.520 the rebuttal phase a response to this, perhaps including some 0:09:07.640 --> 0:09:10.480 of the evidence that might have been left out previously. 0:09:10.800 --> 0:09:15.360 Project Debater didn't really address that criticism. We also found 0:09:15.360 --> 0:09:18.800 out at the end of the debate that typically the 0:09:18.880 --> 0:09:23.839 Intelligence Squared format would include another round the moderator would 0:09:23.840 --> 0:09:26.520 hold around in which you would ask critical questions of 0:09:26.559 --> 0:09:29.480 each of the participants in order to test their arguments 0:09:29.520 --> 0:09:32.640 and their logic. This is done in the normal debates 0:09:32.640 --> 0:09:35.520 on Intelligence Squared, so if you listen to other examples, 0:09:35.559 --> 0:09:38.640 you would hear that round. But Project Debater, while it's 0:09:38.640 --> 0:09:42.000 really impressive, isn't quite up to the task of handling 0:09:42.040 --> 0:09:44.520 that sort of response just yet. And since this was 0:09:44.559 --> 0:09:47.520 really a showcase for the technology, that round was not 0:09:47.600 --> 0:09:51.360 included in this debate. I was really impressed by Project Debater, 0:09:51.800 --> 0:09:54.920 and as Harish pointed out after the exchange, the technology 0:09:54.960 --> 0:09:57.760 has the potential to really help people get a deeper 0:09:57.840 --> 0:10:00.719 understanding of complex topics. They can use it to help 0:10:00.760 --> 0:10:04.439 them support their arguments on any given stance, or and 0:10:04.679 --> 0:10:07.600 this is something I think is equally as important, they 0:10:07.679 --> 0:10:11.320 could use Project Debater to produce counter arguments to their 0:10:11.400 --> 0:10:15.520 own stances. Then they might either better learn how to 0:10:15.640 --> 0:10:19.120 argue against those who oppose them and anticipate the arguments 0:10:19.160 --> 0:10:23.840 they would put up against a specific stance, or it 0:10:23.920 --> 0:10:27.960 might actually change their own mind about the entire subject. 0:10:28.160 --> 0:10:31.240 It might be that you have a preconceived idea of 0:10:31.280 --> 0:10:33.559 what is right, and then you get the information from 0:10:33.559 --> 0:10:35.920 Project Debater and you start to question those ideas. You 0:10:35.960 --> 0:10:38.680 might change your mind. That leads me into the next 0:10:38.679 --> 0:10:42.760 section IBM's general philosophy about artificial intelligence. But before we 0:10:42.800 --> 0:10:53.559 get into that, let's take a quick break. A common 0:10:53.600 --> 0:10:56.840 thread that was once in science fiction and now tends 0:10:56.880 --> 0:10:59.760 to be in today's headlines is the impact that automation 0:11:00.120 --> 0:11:03.640 artificial intelligence will have on the workforce. On the most 0:11:03.679 --> 0:11:07.199 pessimistic side, there's a fear that these technologies are going 0:11:07.240 --> 0:11:10.160 to eliminate millions of jobs and that will be plunged 0:11:10.160 --> 0:11:14.120 into an economic crisis, perhaps requiring an entire overhaul of 0:11:14.120 --> 0:11:17.480 how we think about work and money. And there will 0:11:17.559 --> 0:11:21.199 no doubt be jobs that will become either completely automated 0:11:21.559 --> 0:11:24.079 or automated to the point that fewer humans will be 0:11:24.120 --> 0:11:27.320 needed to carry out that same amount of work over time. 0:11:27.400 --> 0:11:32.520 So there's definitely some validity to that fear, but many entities, 0:11:32.600 --> 0:11:35.599 IBM among them, say that we're probably not going to 0:11:35.679 --> 0:11:40.439 see anything quite so dramatic as a job apocalypse. Instead, 0:11:40.800 --> 0:11:44.200 IBM's vision is one in which artificial intelligence acts sort 0:11:44.200 --> 0:11:48.520 of like a super smart, super efficient assistant to aid 0:11:48.600 --> 0:11:52.360 us in our jobs. The tedious or difficult parts of 0:11:52.480 --> 0:11:56.680 jobs that humans find troubling could be handled by artificial intelligence, 0:11:56.920 --> 0:11:59.640 whereas the parts of jobs that are easy for humans 0:11:59.679 --> 0:12:02.679 but not so easy for machines would still need a 0:12:02.720 --> 0:12:05.960 person taking that position and fulfilling those parts of the 0:12:06.040 --> 0:12:10.480 job duties, and artificial intelligence will necessitate new positions in 0:12:10.559 --> 0:12:13.160 order to oversee the systems and to maintain them and 0:12:13.200 --> 0:12:17.079 grow them over time as businesses themselves grow. I had 0:12:17.120 --> 0:12:19.880 the opportunity to sit down with Rob Thomas, who is 0:12:19.960 --> 0:12:24.079 general manager of IBM Data and AI, to talk about this. Now. 0:12:24.120 --> 0:12:26.680 We're sitting in a lounge in the w Hotel in 0:12:26.720 --> 0:12:29.160 San Francisco for this interview, So if you hear some 0:12:29.240 --> 0:12:32.800 ambient noise that's just the sound of business people being 0:12:32.840 --> 0:12:38.080 business in the background. I'm sitting here with Rob Thomas, 0:12:38.080 --> 0:12:42.600 general manager of IBM Data and AI, and today we 0:12:43.080 --> 0:12:47.360 heard the announcement of Watson Anywhere, and I have to 0:12:47.400 --> 0:12:51.680 ask you what does that mean, Jonathan. It's an exciting 0:12:51.760 --> 0:12:56.000 day for us. Let's start with basics. I like to 0:12:56.040 --> 0:13:00.719 say there's no AI without I, A meaning information in architecture. 0:13:01.679 --> 0:13:04.840 AI is only as good as the data that you 0:13:04.880 --> 0:13:08.760 feed it. So that's a problem every company deals with 0:13:08.920 --> 0:13:11.000 and you can even see it in your consumer life. 0:13:11.000 --> 0:13:12.920 If you're using an app over and over again, it 0:13:12.920 --> 0:13:14.920 starts to know you a little bit. So your AI 0:13:15.040 --> 0:13:17.559 is only as good as your data. What we realized 0:13:17.720 --> 0:13:20.800 is companies have a lot of data, but they have 0:13:20.840 --> 0:13:23.240 a data in a lot of different places. It might 0:13:23.280 --> 0:13:26.240 be in one office location, might be data in a 0:13:26.240 --> 0:13:30.160 different office location. There might be data on a public cloud. 0:13:30.200 --> 0:13:34.280 They might have different cloud providers. We made the decision 0:13:34.360 --> 0:13:37.080 that we were going to bring the AI to the 0:13:37.160 --> 0:13:40.280 data and enable that to happen. So Watson Anywhere is 0:13:40.320 --> 0:13:43.200 about taking the best of what we've built in Watson 0:13:43.559 --> 0:13:45.720 and saying you can have that wherever you want it, 0:13:45.760 --> 0:13:48.920 which is normally wherever your data is. And this is 0:13:48.920 --> 0:13:51.040 going to be significant because this is what clients have 0:13:51.160 --> 0:13:53.600 been asking for and now we're making it really easy 0:13:53.640 --> 0:13:56.280 for them to consume Watson AI wherever they have data. 0:13:56.559 --> 0:13:59.120 So if if I'm understanding this correctly, you can look 0:13:59.120 --> 0:14:02.280 at this in very broad sense in two very different directions. 0:14:02.320 --> 0:14:04.520 You can look at it in the sense of I've 0:14:04.679 --> 0:14:08.200 got this big company and I have data spread out 0:14:08.240 --> 0:14:11.920 through multiple locations, and maybe I need to integrate that 0:14:12.080 --> 0:14:15.280 in meaningful ways. That's one way, but it may also 0:14:15.400 --> 0:14:18.439 mean I have a really large company and I've got 0:14:18.480 --> 0:14:23.960 offices in other states, other countries, perhaps where that integration 0:14:24.080 --> 0:14:27.920 may not be as easy or seamless. There might be 0:14:28.280 --> 0:14:31.000 specific laws, for example, when you need b are over 0:14:31.120 --> 0:14:35.000 in the UK, where I need to be very particular 0:14:35.320 --> 0:14:37.480 with how I'm handling data in this region. I might 0:14:37.520 --> 0:14:39.600 not be applying that somewhere else. In this way, if 0:14:39.640 --> 0:14:42.320 I'm taking the AI to where the data is, I 0:14:42.360 --> 0:14:47.400 can handle those different use case scenarios in the specific 0:14:47.440 --> 0:14:50.280 ways that are necessary. So it could be either way. 0:14:50.320 --> 0:14:52.400 It could be like integrating stuff, or it could be 0:14:52.440 --> 0:14:57.360 applying for specific implementations and I am I a lot 0:14:57.360 --> 0:15:00.720 of companies have different security policies for what they can 0:15:00.720 --> 0:15:03.320 do with their data. Like you say, some are worried 0:15:03.360 --> 0:15:06.040 about g d p R can't leave a certain country. 0:15:06.080 --> 0:15:09.280 So we're just saying take the best of the AI, 0:15:09.440 --> 0:15:11.960 put it wherever it is however you want to do it, 0:15:12.040 --> 0:15:14.600 which makes it really easier for them to access, which 0:15:14.640 --> 0:15:17.720 kind of brings up the idea of So what is Watson? 0:15:17.760 --> 0:15:19.920 What is AI? We can talk about that for a minute. Sure. 0:15:20.680 --> 0:15:23.200 I think there's two worlds of AI right now. One 0:15:23.320 --> 0:15:26.920 is people that want to build their own AI. So 0:15:27.080 --> 0:15:31.000 Watson is a way that you can build run your AI, 0:15:31.360 --> 0:15:34.440 manage that. We have a product called Watson Studio Watson 0:15:34.440 --> 0:15:37.880 Machine Learning that's basically how you build run manager AI. 0:15:38.120 --> 0:15:41.560 That's what the data scientists of the world do. They 0:15:41.600 --> 0:15:44.800 want to build something unique for their company. There's another 0:15:44.840 --> 0:15:47.520 world that says, I don't have those skills to build 0:15:47.520 --> 0:15:51.720 my own AI. I just want to use AI. We've 0:15:51.720 --> 0:15:55.480 built some Watson applications. We have Watson Assistant, which is 0:15:55.480 --> 0:16:00.680 basically a customer service agent encoded in software where we 0:16:00.720 --> 0:16:04.320 automate a lot of the decision making to make current 0:16:04.320 --> 0:16:08.040 customer service representatives a lot more effective and how they 0:16:08.040 --> 0:16:12.360 can support customers. And we've got another application called Watson Discovery. 0:16:12.760 --> 0:16:14.760 Any company with a lot of data in different places, 0:16:14.760 --> 0:16:17.800 they want to discover all of their data, what's in there, 0:16:17.880 --> 0:16:20.800 looking for the proverbial needle in a haystack. So there's 0:16:20.880 --> 0:16:22.440 kind of two worlds of a I I want to 0:16:22.440 --> 0:16:24.880 build my own and then I want to build for me. 0:16:25.240 --> 0:16:27.240 Just help me solve a problem I know I have. 0:16:27.800 --> 0:16:31.080 That's what Watson is today, what we're doing, and so 0:16:31.400 --> 0:16:35.760 you've got the with these two approaches. I also like 0:16:35.840 --> 0:16:38.360 the idea, and it's you sort of alluded to it 0:16:38.400 --> 0:16:43.600 that this AI is really all about augmenting us, not 0:16:43.600 --> 0:16:46.680 not that you're replacing any sort of human element, but 0:16:46.680 --> 0:16:49.720 that you're augmenting what we're already doing, making us more effective, 0:16:49.720 --> 0:16:53.000 more efficient, being able to find more meaning in that data. 0:16:53.600 --> 0:16:55.240 One of the stories we hear all the time is 0:16:55.280 --> 0:16:58.640 about just this concept of big data, this massive amount 0:16:58.640 --> 0:17:03.120 of information that we're constant only uh have at our fingertips, 0:17:03.160 --> 0:17:06.360 but it's so big of a problem that it's hard 0:17:06.400 --> 0:17:10.040 to tackle. This is another approach to doing that in 0:17:10.080 --> 0:17:12.960 a meaningful way, where you're actually able to h to 0:17:13.040 --> 0:17:15.359 create action plans based on all this information. You have 0:17:15.600 --> 0:17:17.320 to say, well, we've got all this info, what do 0:17:17.359 --> 0:17:20.120 we do with it? To me, that's a fascinating part 0:17:20.119 --> 0:17:24.520 of this as well, because obviously one of the boogeymen 0:17:24.920 --> 0:17:28.119 in tech is this concept of of AI. It's usually 0:17:28.320 --> 0:17:32.399 a misunderstanding what AI is, and ibm S approach has 0:17:32.440 --> 0:17:37.359 been No, this is really about enhancement, not about replacement. 0:17:37.880 --> 0:17:41.600 I like to say AI is not going to replace managers, 0:17:42.359 --> 0:17:45.720 but managers that use AI are going to replace managers 0:17:45.720 --> 0:17:49.760 that don't. And it actually it's a wave. It gives 0:17:49.800 --> 0:17:53.080 you a superpower if you're willing to use it. Give 0:17:53.119 --> 0:17:57.280 you an example, Royal Bank of Scotland, big retail commercial bank. 0:17:57.880 --> 0:18:01.040 They're trying to serve all of their ret hell banking customers. 0:18:01.600 --> 0:18:07.200 They're using Watson Assistant for customer service. They're getting faster answers. 0:18:07.960 --> 0:18:10.359 It's still going through their agents, but their agents are 0:18:10.400 --> 0:18:13.359 using the Watson Assistant to say do I understand that question? 0:18:13.440 --> 0:18:16.800 What they're really looking for? So they're getting faster answers, 0:18:16.880 --> 0:18:20.840 they're getting better answers, you get more satisfied clients. So 0:18:21.200 --> 0:18:24.120 that's why I go back to what I said, Managers 0:18:24.119 --> 0:18:28.440 that use AI have superpowers, and so I encourage everybody 0:18:28.480 --> 0:18:31.639 to be open to that because it makes you more effective. 0:18:32.240 --> 0:18:33.760 Means you have to do less of the boring work. 0:18:33.800 --> 0:18:35.480 We all have boring work we have to do. You 0:18:35.520 --> 0:18:37.919 can automate a lot of the boring work. So I 0:18:37.920 --> 0:18:40.520 think it actually makes jobs a lot more interesting, which 0:18:40.560 --> 0:18:45.000 is exciting. And as an average user and average person, 0:18:45.840 --> 0:18:47.600 the one of the results is that you get you 0:18:47.680 --> 0:18:50.840 just get better results when you're using these different services 0:18:50.920 --> 0:18:53.920 that are incorporating the AI, because you're getting the right answer, 0:18:53.960 --> 0:18:56.920 You're getting the right answer faster. Uh, you don't have 0:18:57.000 --> 0:19:00.200 to worry about as much follow up. So it's takes 0:19:00.200 --> 0:19:03.280 a lot of the frustration out of those interactions between 0:19:03.760 --> 0:19:07.959 customer and say, uh, you know a customer representative. I 0:19:08.000 --> 0:19:10.760 think I've been on both sides of that. I've been 0:19:10.760 --> 0:19:13.920 on the side as the customer who's frustrated, trying very 0:19:14.000 --> 0:19:16.399 hard not to let my frustration spill out with my 0:19:16.480 --> 0:19:20.720 interaction representative. And I've I've married to a woman who 0:19:20.840 --> 0:19:24.119 was a customer representative who would come home and I 0:19:24.160 --> 0:19:26.520 would I would hold her for an hour because she 0:19:26.600 --> 0:19:30.359 had been nailed at eight hours straight. So uh, that 0:19:30.480 --> 0:19:32.240 is something I think that a lot of people lose 0:19:32.280 --> 0:19:34.359 in this too, because when we have these discussions, we're 0:19:34.400 --> 0:19:38.399 talking about the enterprise level frequently, and the average person says, well, 0:19:38.400 --> 0:19:42.160 how does that affect me? It affects them because this 0:19:42.280 --> 0:19:46.000 ends up being incorporated into applications that are forward facing 0:19:46.000 --> 0:19:51.080 for customers in some cases. So I'm really excited by this. 0:19:51.400 --> 0:19:56.119 I think Watson is an incredible platform. I've had various 0:19:56.160 --> 0:19:59.960 interactions on various levels with Watson throughout the years, ray 0:20:00.040 --> 0:20:03.679 raging from seeing how it could be used in a 0:20:03.720 --> 0:20:07.440 customer service aspect to Watson Chef, which was my favorite 0:20:07.440 --> 0:20:12.000 implementation I've ever seen, just to get weird fun recipes 0:20:12.200 --> 0:20:17.480 generated by Watson Um. Last night I saw the debater presentation, 0:20:17.480 --> 0:20:20.919 which was fascinating to see this platform be able to 0:20:20.960 --> 0:20:25.439 actually put together a cohesive, coherent argument from lots of 0:20:25.480 --> 0:20:30.040 different points of data. That is a phenomenal achievement. People 0:20:30.080 --> 0:20:33.679 don't realize how incredibly difficult that is. What excites you 0:20:33.920 --> 0:20:36.960 most about where we are with AI and where you 0:20:37.000 --> 0:20:39.920 see us going in the future. Let me give you 0:20:39.960 --> 0:20:42.800 an example that I think hopefully everybody can relate to, 0:20:43.080 --> 0:20:44.840 which will kind of bring it to life for how 0:20:45.440 --> 0:20:50.320 AI is not just impacting businesses but individuals. AMC Networks 0:20:50.440 --> 0:20:53.199 is a big client of ours, so I'm sure some 0:20:53.280 --> 0:20:55.919 of your listeners have seen some of their TV shows. 0:20:55.960 --> 0:20:59.000 Breaking Bad was a popular one, but they've got many 0:20:59.240 --> 0:21:04.560 and AMC networks challenge was, how do we understand what 0:21:04.760 --> 0:21:09.840 viewers are responding to, what are they liking? And can 0:21:09.880 --> 0:21:12.879 we adjust plot lines based on what we're hearing and 0:21:12.960 --> 0:21:16.320 what they're liking. They also, you know, their businesses also 0:21:16.320 --> 0:21:19.959 around advertising, so can we give advertisers an idea of 0:21:19.960 --> 0:21:22.520 when to engage with the users, how to engage with 0:21:22.520 --> 0:21:27.440 the users. They're using our technology on cloud where they 0:21:27.640 --> 0:21:31.120 federate data coming from step top boxes companies like Nielsen 0:21:31.200 --> 0:21:35.720 third party data. They bring that behind their firewall and 0:21:35.760 --> 0:21:37.720 the company to manage that data. And then they're using 0:21:37.760 --> 0:21:41.120 things like Watson Studio to build models and say hey, 0:21:41.200 --> 0:21:43.760 you should go do this kind of thing to an advertiser, 0:21:44.280 --> 0:21:46.840 or even reaching out direct to a consumer saying we 0:21:46.920 --> 0:21:50.840 thought this show might interest you. So AI is behind 0:21:50.920 --> 0:21:55.920 the scenes actually everywhere, and I think sometimes you only 0:21:55.960 --> 0:21:58.000 notice that when you kind of have an aha moment 0:21:58.040 --> 0:22:00.600 where you're like, wow, that felt magical. It's it's actually 0:22:00.640 --> 0:22:03.600 AI is not magic, it's just computer science. But it 0:22:03.720 --> 0:22:07.840 is impacting every individual, whether they know it or not. Today. 0:22:07.920 --> 0:22:11.000 That's incredible. I only have one last question for you, 0:22:11.400 --> 0:22:13.679 which is are there any questions I should have asked you? 0:22:14.520 --> 0:22:17.840 I wish you would ask me why IBM and how 0:22:17.960 --> 0:22:21.760 is how is IBM relevant to this space going forward? 0:22:22.000 --> 0:22:24.960 I'm sorry I did have one more question. Okay, why 0:22:25.000 --> 0:22:28.480 IBM and why is IBM relevant in this space going forward? 0:22:28.640 --> 0:22:32.879 Great question. IBM has an amazing history over a hundred 0:22:32.920 --> 0:22:36.240 years old, and I think we've always been a steward 0:22:36.560 --> 0:22:41.600 of responsibility and integrity. And when you work with IBM, 0:22:41.760 --> 0:22:43.280 you know what you're going to get, which is you're 0:22:43.280 --> 0:22:47.800 going to be satisfied. This whole area of data AI 0:22:48.119 --> 0:22:52.680 make some people a little squeamish. They're worried about lack 0:22:52.760 --> 0:22:56.440 of transparency. They're worried about is my data being shared? 0:22:57.280 --> 0:22:59.600 The best part of working with IBM and being part 0:22:59.640 --> 0:23:02.879 of him is that you know your data safe, you 0:23:02.920 --> 0:23:05.480 know your models are safe. You know IBM not sharing 0:23:05.520 --> 0:23:07.760 this with anybody else. You know that we will be 0:23:07.800 --> 0:23:11.240 the stewards of responsibility and AI. That's why last year 0:23:11.240 --> 0:23:14.639 we came out with explainability and bias detection for AI. 0:23:14.760 --> 0:23:16.800 I think we're the first company to do that, because 0:23:17.040 --> 0:23:19.320 there's a lot of things that can go wrong in 0:23:19.359 --> 0:23:23.119 the world of machine learning or AI unless you're thinking 0:23:23.160 --> 0:23:27.560 about societal impacts, human impacts, and we spend a lot 0:23:27.560 --> 0:23:29.400 of time on that at IBM. So that's why I'm 0:23:29.480 --> 0:23:33.280 very optimistic. Absolutely, and yes, you definitely don't want something 0:23:33.280 --> 0:23:36.520 like artificial intelligence to become a black box technology where 0:23:36.520 --> 0:23:39.120 you have no idea how it's making its decisions behind 0:23:39.160 --> 0:23:43.920 the scenes, because obviously that breeds mistrust and unease. So 0:23:44.040 --> 0:23:46.640