AI, China, United States discussed on WSJ The Future of Everything

Automatic TRANSCRIPT

Are you hiring with indeed you can post a job in minutes? Set up screener questions then zero in on your shortlist of qualified candidates using an online dashboard get started today at indeed dot com slash podcast. That's indeed dot com slash podcast. What is our decision intelligence by this? I don't mean singularity super intelligence that is way off not worth talking about. I'm talking about the use of machine learning technologies in the single domain with a large amount of training. Data reaching superhuman performance. Kaifu Lee is one of China's most influential tech figures and venture capitalists. He spoke to us about artificial intelligence at our tech conference de live in November thing about Amazon taking a lot of clicks and buys it into knowing what you will by knowing what you will buy more than, you know, yourself from search results to music playlist to digital assistance like Siri, and Alexa is more embedded in our everyday lives than ever before. But it's fine. No means universal. And it's not good at everything at least not yet. So the current state of the art is something narrow. Dr IP park is the president of L G. I spoke to him in January Kirk AI can do something extremely well in narrow area. So he can pay chest can be chess champion you can pay good game. But if you ask that Chesler game to say go for shopping for you something to do that because it's so narrowly focused the field has to sort of evolve into a general, that's where real intelligible come in as data feeds the advancement of machine learning and other AI technologies their value in business is surging as an industry. Hey, I is expected to generate trillions and additional wealth for the world over the next decade companies from TD Bank to apple have expanded their C suites to include chief AI officers tasked with understanding the benefits and biases. It can bring to a business. Now, the US is boosting spending on AI related research in what's seen is a move to counter China's growing prowess in the field from the Wall Street Journal. This is the future of everything. I'm Jennifer strong. History shows when there's Grayton technology inventions, physical infrastructure in societal infrastructure has to change. Ginny remedy is the chief executive of IBM, she says companies need to understand how AI can augment their strategies or risk being left behind. She spoke with the Wall Street Journal's editor in chief Matt Murray at the World Economic Forum in Davos, Switzerland. When you talk about when you think about what does it mean to you explain to us? What is what it isn't do people understand fully what's coming our way today? They misunderstand. How do you define help us think about it? I think if it is simply as a way to make each of us better. I think if the boils down to one thing it's about making better decisions. And in that is in fact, it's one of our values that we've really consciously stated that the purpose of these technologies is to augment man, not replace man. And I think those of us who build them have allowed to do. With that what it is. So I think that when I think of it, and that's why actually for business. We'll kind of go between why do people feel like you just described in for what I see in clients. What is everyone doing with AI because I'm actually starting to see much greater deployment now? And I can talk about what people learned and what didn't go so well, and what has gone well. But basically if you kind of think back in time would go back to areas of IT in. There was a big advantage given to processing and back all our back offices changed. Yours included all changed, then there's been this most recent era, I think everybody would talk about networking networking, neck, the platforms all came up. And now people are going to compete at expertise. And so that's this is one way to help build expertise not just in a company, but a process in anything. So that's what I think of it as client facing producer, internal expertise wells. And that's a really important point. Actually the first in stanch. Station. Most AI today is all customer service. So every client I've met with here, and those that are far ahead have gone well beyond that those beginning there, we say, what's what's the first area was everything, but he's doing, and I was telling the very first thing is customer service the sleeper area, which is a way to get a whole company to understand. Why is HR in most people are very surprised about that? Right. That that's a way. But I wanna can I just park that for a sec. So back to your point of been why most people don't even think of HR having. I little. Oh, that isn't I feel differently about my? The. I actually think most most companies would tell you they've underinvested in that area by far right in. So it can actually be an area that and it's going to be related to the topic. You said why do people feel right now? This is that's what you feel in the air here. And I think there's two reasons one is I would call it a trust headwind. And then the second reason is I said, it's job crisis. But I think I would be better to reframe it as it's a large issue of inclusivity or people feeling left behind in the future. So if I could just maybe just talk a little bit about each one of them on the trust headwind. I think that's out there. You've got people. They know all the state is there. They know there's personal data out there this concern about what is it going to be used for is it going to be replaced is it unfair advantage against me. And I feel very strongly. There's two things that have to happen on trust one is not just an IT company. Matt every company should be clear. What they're. Really clear what their principles are about this data, and then be willing to be audited in accountable to them this goes back 'cause we've been had a long time actually was it was year before I was born was some of this beginning of the work is done. It's been a very long time, but it's gone through its winters and come back out, but our principles our first one is what we're gonna do the purpose of AI in. These new technologies is dogmatic man, I said that one second one is really clear that data belongs to its owner. And it's insights belong to the owner. Third thing is it's gotta be explainable and transparent all these technologies. So kind of asking what regulation I was going to talk about you, you might be headed where I was going. But but that makes sense. But I think that when when you hear that from IBM IBM's got you Facebook has of you Microsoft as if you give us a good or bad. I mean fully was you're in one of the things he talks about as you know, in his book is that part of the Chinese advantage that he sees. Is data and China has a model, and the China model is they've got your data. They do all kinds of things with the data Europe has a different model with GDP are. And again this year. I think the feeling is your may miss out on technology because of that I'm not sure that the United States yet has a clear common, Sandra. If I'm not sure if the government needs to set one or companies need to do it together, but from company to company and interaction their action. I think that's part of the exile. We don't know how you feel. Yeah. I think it I think you're conflicting two different things that don't necessarily have to be completed one is that you must have massive amounts of AI a data to have a because the most most interesting things we are working on now. And it's more than just working on is things like one shot learning where it is not massive amounts of data. And so that is where the field is heading that it will not be just massive amounts, of example of. Oh, so it's very simple. Yes, here here's one that. We're working on right now. I'm an insurance company and. And because the biggest inhibitor of getting a I in the training. It's eighty five percent of any effort. Client does is. I'm getting the data getting it ready and not on the actual algorithms. And so I'm doing claims for roofs hail coming down on roofs. Right. I've all my pictures roofs done after Trane roof roof after roof one shot learning would be this roost been damaged by hail. I show a picture of one roof with hail on it. And now it knows what all the other kind of roofs could look like if they had hail out them. And that's a real one. We're working on with insurance company. That's one shot learning versus. I gotta retrain on. What does hey, look like on every different kind of structure that there could be and that is a that to me is a huge huge barrier removed. If you can do that, you will get widespread. A I in business because why time to value and costs will be way less to do it. So so that's why wouldn't necessarily I understand that argument. But I do think one that we do have to come together on is when you think about what you say it's easy for all of us to say are principles, by the way. Everyone has not been clear what they're. Simples are. But particularly when it comes to the consumer. I do think there needs to be very precise regulation it because my big fear. One of the things I've talked about here is if all governments are feeling pressure to do something on this in its particularly consumer privacy. And so what what is a to me common sense, but targeted regulation would be for consumer. You should be able to know what data's being collected opt out get it removed. See what you have as minimum, right and simple points. There should be liability for illegal things. Right. I think there should be in get to agreement on those things because otherwise my biggest fear is governments without understanding. All this will go too overboard in actually really really impacted digital economy. Is there a prospect of the kind of regulation talking about it in the United States right now, very hard. If it's really come together. Yeah. No. I think there is some prospect of that, right? Because in its born of people hearing all different other kind that has unintended consequences. So I just when you put those two things together about you know, is is one country ahead because it has all the data. I don't necessarily this is going to be raised about who has it necessarily just all the data. You've said one hundred percent of jobs are going to be changed dot dot very distant future. Yeah. So you're not dimissed that's queer. My when you talk about the things, I think one of the things that at least I hear there's an actual logic to some of these things start as processes, but obviously when you talk about something like extending the roof thing that does have implications. So there's going to be I assume he lot of job displacement.

Coming up next