WEBVTT - IBM Chairman & CEO Arvind Krishna Talks AI Growth, Regulation

0:00:00.120 --> 0:00:04.280
<v Speaker 1>We're focusing on how IBM is advancing generative AI. Arvind

0:00:04.360 --> 0:00:07.080
<v Speaker 1>Krishna he is IBM chairman and CEO. He spoke with

0:00:07.160 --> 0:00:10.280
<v Speaker 1>Wall Street Week host David Weston yesterday at the Milkin

0:00:10.360 --> 0:00:12.799
<v Speaker 1>Institute Global Conference in Los Angeles.

0:00:15.120 --> 0:00:17.239
<v Speaker 2>Exactly a year. I think if edymon but I David.

0:00:17.239 --> 0:00:20.680
<v Speaker 3>We announced it what's the next product at think in

0:00:20.760 --> 0:00:24.000
<v Speaker 3>May last year, and since then, the excitement in our

0:00:24.040 --> 0:00:29.320
<v Speaker 3>clients embracing the technology, getting the technology deployed, asking for expertise,

0:00:29.360 --> 0:00:30.560
<v Speaker 3>and getting projects going.

0:00:30.800 --> 0:00:32.360
<v Speaker 2>I think all of that has been wonderful.

0:00:32.880 --> 0:00:35.560
<v Speaker 3>Our inception to date book of business nine seeds of

0:00:35.640 --> 0:00:36.400
<v Speaker 3>billion dollars.

0:00:36.640 --> 0:00:38.720
<v Speaker 2>That's pretty good for a year. And I think and

0:00:38.760 --> 0:00:39.960
<v Speaker 2>it speaks to the excitement.

0:00:40.240 --> 0:00:42.640
<v Speaker 3>And I think what's even more, our clients are excited

0:00:42.640 --> 0:00:44.760
<v Speaker 3>about what it does for them. You can see them

0:00:45.000 --> 0:00:49.360
<v Speaker 3>the user for customer experience, for programming apporting as it's called,

0:00:49.680 --> 0:00:52.640
<v Speaker 3>as well as for helping improve the enterprise. All of

0:00:52.680 --> 0:00:56.120
<v Speaker 3>this put together is really exciting in how our enterprise

0:00:56.160 --> 0:00:57.880
<v Speaker 3>clients are embracing the generative AI.

0:00:58.200 --> 0:01:00.360
<v Speaker 4>One of the things that people anticipate from general it's

0:01:00.360 --> 0:01:04.440
<v Speaker 4>an increase in productivity. Are you finding that with your customers.

0:01:04.520 --> 0:01:07.039
<v Speaker 4>Are they increasing criticulity and how do you measure that

0:01:07.120 --> 0:01:08.440
<v Speaker 4>increase in productivity?

0:01:08.840 --> 0:01:10.840
<v Speaker 2>I think David is both.

0:01:11.240 --> 0:01:14.520
<v Speaker 3>I think increase in productivity is important, but productivity can

0:01:14.560 --> 0:01:17.360
<v Speaker 3>be made into just cost cutting our efficiency. That's not

0:01:17.440 --> 0:01:20.240
<v Speaker 3>the primary reason. The primary reason I think our clients

0:01:20.240 --> 0:01:23.679
<v Speaker 3>are excited. Can they get more business done while holding

0:01:23.720 --> 0:01:26.880
<v Speaker 3>their costs in somewhat in control? And so it's really

0:01:26.959 --> 0:01:31.920
<v Speaker 3>about more of a revenue generator then productivity alone. Now,

0:01:31.920 --> 0:01:34.240
<v Speaker 3>of course it is a revenue generator because it's making

0:01:34.280 --> 0:01:36.560
<v Speaker 3>you more productive, and I think that is what is exciting.

0:01:36.720 --> 0:01:38.160
<v Speaker 2>Think about customer experience.

0:01:38.480 --> 0:01:40.720
<v Speaker 3>Yeah, you may save a dollar on somebody's time and

0:01:40.760 --> 0:01:43.560
<v Speaker 3>a call center, but that's not why. If you can

0:01:43.600 --> 0:01:46.679
<v Speaker 3>make the end client more satisfied they got that answer quicker,

0:01:46.840 --> 0:01:49.160
<v Speaker 3>they've got a better answer. Are they more likely to

0:01:49.200 --> 0:01:51.680
<v Speaker 3>come back and do more repeat business? I think that's

0:01:51.760 --> 0:01:52.720
<v Speaker 3>far more exciting.

0:01:53.520 --> 0:01:56.040
<v Speaker 4>There's a lot of discussion about how general AI will

0:01:56.080 --> 0:01:59.880
<v Speaker 4>grow and how it will evolve. You've really staked your

0:02:00.080 --> 0:02:04.480
<v Speaker 4>claim on an open architecture, including for IBM, but also

0:02:04.600 --> 0:02:08.240
<v Speaker 4>with Meta in your AI alliance. Tell us why that's important.

0:02:08.240 --> 0:02:10.639
<v Speaker 4>Why is open architecture important in AI?

0:02:11.200 --> 0:02:13.960
<v Speaker 2>Look, it's always around how do you drive innovation?

0:02:14.600 --> 0:02:16.800
<v Speaker 3>And in a case where people worry a lot about

0:02:16.880 --> 0:02:19.680
<v Speaker 3>what is the fair use, what is the copyright, what

0:02:19.760 --> 0:02:22.639
<v Speaker 3>is the IP protection of the data and methodology used

0:02:22.639 --> 0:02:24.920
<v Speaker 3>to train AI, you have both those going on.

0:02:25.400 --> 0:02:27.400
<v Speaker 2>So in order to drive open.

0:02:27.360 --> 0:02:29.320
<v Speaker 3>What we are doing is we're taking some of our

0:02:29.360 --> 0:02:32.680
<v Speaker 3>base models, as Meta has done, and we are putting

0:02:32.680 --> 0:02:36.040
<v Speaker 3>them out under an open license for the epaty license,

0:02:36.240 --> 0:02:38.200
<v Speaker 3>which means people are free to build upon it.

0:02:38.480 --> 0:02:40.239
<v Speaker 2>And what the build upon is their. They don't have

0:02:40.320 --> 0:02:41.200
<v Speaker 2>to give it back to us.

0:02:41.720 --> 0:02:44.280
<v Speaker 3>When you care a lot about maybe some proprietary data

0:02:44.320 --> 0:02:47.400
<v Speaker 3>instead an enterprise and you use that AD skills to

0:02:47.480 --> 0:02:50.280
<v Speaker 3>an AI model and you can now say, oh, because

0:02:50.320 --> 0:02:52.920
<v Speaker 3>yours is open, it's now mine what is best?

0:02:53.120 --> 0:02:56.400
<v Speaker 2>I think it makes it block got interesting to the enterprise.

0:02:56.600 --> 0:03:00.519
<v Speaker 4>So IBM with wats an excess really pursuing enterprise base AI.

0:03:01.120 --> 0:03:05.520
<v Speaker 4>Does that get around the challenges of intellectual property because

0:03:05.520 --> 0:03:09.800
<v Speaker 4>there are issues in consumer facing about copyright about what

0:03:09.800 --> 0:03:11.560
<v Speaker 4>you're going to learn on? Is it the way it

0:03:11.600 --> 0:03:14.560
<v Speaker 4>works for you? The customers using their own data to

0:03:14.680 --> 0:03:16.399
<v Speaker 4>actually educate the model.

0:03:16.520 --> 0:03:19.800
<v Speaker 3>Absolutely, and we're just putting out you talked about open

0:03:20.000 --> 0:03:24.040
<v Speaker 3>right now, we're also right today and tomorrow putting out

0:03:24.040 --> 0:03:28.520
<v Speaker 3>a new technology, a really innovative technology called instruct lab.

0:03:28.960 --> 0:03:31.639
<v Speaker 3>This allows a client to take one of us or

0:03:31.639 --> 0:03:35.360
<v Speaker 3>maybe somebody else's model, take it inside that enterprise, train

0:03:35.440 --> 0:03:37.920
<v Speaker 3>it with some data and some methodology of their own,

0:03:38.240 --> 0:03:40.960
<v Speaker 3>and then the layer on top inside instruct lab is this.

0:03:41.520 --> 0:03:43.440
<v Speaker 2>It does not need to come back to anybody else.

0:03:43.800 --> 0:03:46.920
<v Speaker 2>I think that gets rid of the whole question of

0:03:47.000 --> 0:03:48.560
<v Speaker 2>copyright and fair use.

0:03:49.320 --> 0:03:53.120
<v Speaker 4>How does AI alignes fit with possible government regulation. We

0:03:53.200 --> 0:03:55.360
<v Speaker 4>have some moves in Europe, as you know, there's an

0:03:55.360 --> 0:03:58.640
<v Speaker 4>executive order in the United States divide administration is put out.

0:03:58.800 --> 0:04:01.080
<v Speaker 4>How do those two things fit? Because you talk about

0:04:01.080 --> 0:04:03.920
<v Speaker 4>standards as part of the AI Alliance.

0:04:03.840 --> 0:04:07.160
<v Speaker 3>I think every regulator is worried about three topics, not

0:04:07.360 --> 0:04:11.240
<v Speaker 3>just safety in regulation. They're wanted about innovation. They're worried

0:04:11.280 --> 0:04:14.440
<v Speaker 3>about competition, and they're worried about safety and regulation.

0:04:14.720 --> 0:04:17.600
<v Speaker 2>So when you take those three together, the AI Alliance

0:04:17.640 --> 0:04:20.840
<v Speaker 2>that open really come together to help you foment innovation.

0:04:21.240 --> 0:04:23.560
<v Speaker 3>So I think that that actually helps the regulators to

0:04:23.640 --> 0:04:26.880
<v Speaker 3>think about what is going on here. While I in caution,

0:04:27.360 --> 0:04:29.960
<v Speaker 3>there will be some gardens that are always fut but

0:04:30.120 --> 0:04:33.640
<v Speaker 3>in my experience, open technologies have always been safer and

0:04:33.720 --> 0:04:35.920
<v Speaker 3>more secure than close technologies.

0:04:36.320 --> 0:04:38.719
<v Speaker 4>Is one of the risks that maybe you're obviating with

0:04:38.800 --> 0:04:41.240
<v Speaker 4>your emphasis on open architecture. That's some of that. I'll

0:04:41.240 --> 0:04:43.680
<v Speaker 4>call them big guys get an advantage and really have

0:04:43.760 --> 0:04:44.640
<v Speaker 4>an entrenched position.

0:04:45.520 --> 0:04:47.680
<v Speaker 2>Well, I'll use the concert of a wall garden.

0:04:49.080 --> 0:04:53.120
<v Speaker 3>When you have a walled garden, has those areas and

0:04:53.160 --> 0:04:55.440
<v Speaker 3>technologies been more innovative.

0:04:55.000 --> 0:04:55.880
<v Speaker 2>Or less innovative?

0:04:56.279 --> 0:04:58.000
<v Speaker 3>All of a sudden, the wall garden has always been

0:04:58.080 --> 0:05:01.239
<v Speaker 3>less innovative, and so I think that it actually helps

0:05:01.279 --> 0:05:05.440
<v Speaker 3>you create more competition. Does it avoid regulatory lock in

0:05:05.640 --> 0:05:07.880
<v Speaker 3>off a certain one or two players?

0:05:08.240 --> 0:05:10.279
<v Speaker 2>Likely? But isn't that good for all of us?

0:05:10.800 --> 0:05:12.680
<v Speaker 4>Are you pro regulation?

0:05:13.520 --> 0:05:16.080
<v Speaker 3>I am pro regulation as long as is it as

0:05:16.160 --> 0:05:19.320
<v Speaker 3>light touch and allows innovation to happen.

0:05:19.680 --> 0:05:21.360
<v Speaker 2>I absolutely would be pro regulation.

0:05:21.760 --> 0:05:32.240
<v Speaker 3>If regulation tries to reduce innovation, I think that's a problem.

0:05:28.920 --> 0:05:31.159
<v Speaker 1>And that, of course was Arvind Krishna. He is IBM

0:05:31.279 --> 0:05:34.520
<v Speaker 1>Chairman and CEO and Wall Street Week hoast David Wesson