1 00:00:00,040 --> 00:00:04,040 Today we're talking about generative AI and leadership 2 00:00:04,040 --> 00:00:08,680 with Paul Doherty, the Global Chief Executive for Accenture 3 00:00:08,680 --> 00:00:13,080 Technology. My guest cohost is Q Harrison 4 00:00:13,080 --> 00:00:17,560 Terry, the Chief Growth Officer for the Mark Cuban Companies. 5 00:00:18,000 --> 00:00:21,040 Thanks for having me, Mike, and it's exciting to be able to talk 6 00:00:21,040 --> 00:00:25,220 with Paul on AI today. Paul, why don't we begin by 7 00:00:25,220 --> 00:00:29,380 asking you to tell us about your work as the Chief Executive for 8 00:00:29,380 --> 00:00:31,300 Technology at Accenture? Yeah. 9 00:00:31,300 --> 00:00:33,700 Accenture is a large organization. 10 00:00:33,700 --> 00:00:38,420 We're about 740,000 people over $60 billion in revenue. 11 00:00:38,990 --> 00:00:42,510 And we help companies do amazing things with technology, You 12 00:00:42,510 --> 00:00:43,470 know, that's what we're all about. 13 00:00:43,470 --> 00:00:48,430 Do you want to give us to start a just kind of a brief overview 14 00:00:48,790 --> 00:00:50,870 of generative AI? I think everybody in the 15 00:00:50,870 --> 00:00:55,310 audience knows what it is, but in the context of business and 16 00:00:55,310 --> 00:00:57,950 in our world, where does it fit today? 17 00:00:58,270 --> 00:01:00,990 Talk about generative AI. You have to talk about AI 1st, 18 00:01:01,030 --> 00:01:06,030 and AI has been around for a long time and all of us use AI 19 00:01:06,030 --> 00:01:08,110 continuously. You know anybody you know? 20 00:01:09,090 --> 00:01:11,410 Those of us, the three of us talking here and anybody 21 00:01:11,410 --> 00:01:15,010 listening has used a a I dozens, if not hundreds of times today. 22 00:01:15,010 --> 00:01:17,890 So it's because, you know, a I has become a pervasive part of 23 00:01:17,890 --> 00:01:20,130 our life through the advances in machine learning and deep 24 00:01:20,130 --> 00:01:22,370 learning and such that have come before. 25 00:01:22,770 --> 00:01:26,490 And a I, as I'm sure you know, most of the audience knows it's 26 00:01:26,490 --> 00:01:30,330 an old field if the term was invented, I believe in 1953 at a 27 00:01:30,330 --> 00:01:32,610 conference in Dartmouth 70 years ago. 28 00:01:32,810 --> 00:01:35,250 And it's gone through a lot of iterations over the years. 29 00:01:35,690 --> 00:01:38,650 So we I like to think about 3 forms of a I. 30 00:01:38,930 --> 00:01:43,530 Diagnostic A I which is using a I to diagnose things often you 31 00:01:43,530 --> 00:01:46,890 know deep learning and the like to look at you know for example 32 00:01:46,890 --> 00:01:50,250 using machine vision to look for manufacturing defects. 33 00:01:50,250 --> 00:01:52,370 The other thing we do commonly or to unlock our phones, you 34 00:01:52,370 --> 00:01:55,850 know as we do, as we do every every few minutes of every day 35 00:01:56,010 --> 00:01:59,010 or assisted driving features in cars and then there's that's 36 00:01:59,010 --> 00:02:00,370 diagnostic and then there's predictive. 37 00:02:00,370 --> 00:02:04,450 A I such as a I we use to do retail forecasting for companies 38 00:02:04,450 --> 00:02:07,410 often you know machine learning and optimization models those 39 00:02:07,410 --> 00:02:10,410 are well established techniques. So we have, you know, lots of 40 00:02:10,410 --> 00:02:13,010 people doing that work for lots of clouds around the world and 41 00:02:13,090 --> 00:02:16,250 many, many companies use it. Generative A I now is the new 42 00:02:16,250 --> 00:02:19,170 thing on the scene and it really is a massive breakthrough, 43 00:02:19,170 --> 00:02:21,610 probably the biggest breakthrough in A I today. 44 00:02:22,410 --> 00:02:25,250 And what we're really talking about with generative A I is 45 00:02:25,250 --> 00:02:27,890 foundation models, which are really powerful models. 46 00:02:28,450 --> 00:02:32,090 That can be reused in across many different use cases. 47 00:02:32,090 --> 00:02:33,690 That's why they're called foundation models. 48 00:02:33,890 --> 00:02:37,610 Large language models are are type of foundation models that 49 00:02:37,770 --> 00:02:41,210 that really understand language and have mastered have allowed 50 00:02:41,210 --> 00:02:44,570 us to really master language through artificial intelligence 51 00:02:44,890 --> 00:02:47,450 and then the transformer technology added on top of that 52 00:02:47,450 --> 00:02:50,290 allows us to to generate things. That's why, you know GPT is 53 00:02:50,290 --> 00:02:54,490 generative pretrain transformer. It's the it's these large models 54 00:02:54,690 --> 00:02:58,370 that then have transformer technologies they can create. 55 00:02:58,650 --> 00:03:00,490 New sources of content. So that's really the 56 00:03:00,490 --> 00:03:03,170 breakthrough of generative A, I models that are very powerful 57 00:03:03,170 --> 00:03:05,890 and can be reused rather than bespoke data science projects 58 00:03:06,130 --> 00:03:09,530 combined with foundation models which have tremendous reuse and 59 00:03:09,530 --> 00:03:12,690 power combined with this creative capability to produce 60 00:03:12,690 --> 00:03:16,490 language content, whether it be graphics, video, etcetera. 61 00:03:16,890 --> 00:03:20,810 And it really is really transformational in terms of 62 00:03:20,810 --> 00:03:23,770 what it allows us to do as it is individuals and what it allows 63 00:03:23,770 --> 00:03:26,010 companies to do. But we're at the very early 64 00:03:26,010 --> 00:03:27,820 stages still. Hey, Paul. 65 00:03:27,820 --> 00:03:30,780 One of the things that I want to talk to you about today is the 66 00:03:30,780 --> 00:03:34,020 whole concept of you thinking about this stuff. 67 00:03:34,020 --> 00:03:38,460 I mean, almost a decade ago in your book, Human plus Machine 68 00:03:38,860 --> 00:03:40,940 reimagining work in the Age of AI. 69 00:03:40,940 --> 00:03:43,380 I don't, sorry, I don't have it in front of me, but I did read 70 00:03:43,380 --> 00:03:46,220 it a while ago. And when I was looking back at 71 00:03:46,220 --> 00:03:49,500 that book, one of the things that you talked about was how AI 72 00:03:49,500 --> 00:03:52,980 would ultimately become the ultimate innovation machine. 73 00:03:53,340 --> 00:03:58,980 And it's it's fascinating that it's 2023, a few, almost five 74 00:03:58,980 --> 00:04:00,820 years later since you published that book. 75 00:04:01,780 --> 00:04:04,540 What's your take like that? It seems like you were, it seems 76 00:04:04,540 --> 00:04:07,740 like you're spot on, but what things happen in generative AI 77 00:04:07,740 --> 00:04:11,900 that you didn't envision or forecast back in 2018? 78 00:04:12,470 --> 00:04:15,910 I think that the premise and the all the precepts in human plus 79 00:04:15,910 --> 00:04:20,510 machine really, really have stood up the the stood the test 80 00:04:20,510 --> 00:04:23,390 of time well and the concepts we talked about the human plus 81 00:04:23,390 --> 00:04:27,110 machine and the idea that a I gives humans superpowers to do 82 00:04:27,110 --> 00:04:30,110 new things really has you know stood the test of time. 83 00:04:30,110 --> 00:04:32,990 And we see generative A I as an even bigger step forward in 84 00:04:32,990 --> 00:04:35,830 terms of you know, the the augmentation and enhancement of 85 00:04:35,830 --> 00:04:38,990 what it can do, you know for for all of us in terms of giving us 86 00:04:38,990 --> 00:04:41,590 the greater tools and productivity to do, you know, to 87 00:04:41,590 --> 00:04:44,230 do new things. I think that the surprise we did 88 00:04:44,230 --> 00:04:47,070 talk about, you know, all this technology in that book and then 89 00:04:47,070 --> 00:04:49,590 our next book. And my coauthor and I wrote Jim 90 00:04:49,590 --> 00:04:52,270 Wilson's, my coauthor, which was called Radically Human, that was 91 00:04:52,270 --> 00:04:54,710 the 2nd book. But it did. 92 00:04:54,830 --> 00:04:57,670 The pace of the advance is what surprised us more so than the 93 00:04:57,670 --> 00:05:00,070 capability. We're anticipating that some of 94 00:05:00,070 --> 00:05:03,990 these capabilities would come along, but with the, you know, 95 00:05:03,990 --> 00:05:06,870 the pace of development of the foundation models, the rapid 96 00:05:07,190 --> 00:05:11,250 growth as. You know the size and complexity 97 00:05:11,250 --> 00:05:13,730 of the parameters and the weightings and everything and 98 00:05:13,890 --> 00:05:15,650 you know the breakthroughs that came about with that. 99 00:05:15,650 --> 00:05:18,090 We're probably the biggest surprise you know queue in terms 100 00:05:18,090 --> 00:05:20,810 of what we saw and. Then one last thing on that is 101 00:05:20,810 --> 00:05:23,730 when you talk about the timing and how like just fast 102 00:05:23,730 --> 00:05:26,770 everything is is coming together, it's it's fascinating 103 00:05:26,770 --> 00:05:30,890 to think that you know even Open a Eyes ChatGPT is still 104 00:05:30,890 --> 00:05:34,410 relatively like 9 to 10 months old as we stand today. 105 00:05:34,900 --> 00:05:38,860 And then when we Fast forward I mean like to just yesterday Elon 106 00:05:38,860 --> 00:05:43,060 Musk announced ex A I which is another you know fascinating A I 107 00:05:43,060 --> 00:05:48,140 company as a business leader and executive how should I think 108 00:05:48,140 --> 00:05:51,540 about A I I mean it's happening fast but does that mean I take 109 00:05:51,700 --> 00:05:55,740 you know the move fast and break things approach or should I wait 110 00:05:55,740 --> 00:06:01,380 and see where things settle with our but but but on the flip side 111 00:06:01,380 --> 00:06:03,940 of that our organization might you know be behind. 112 00:06:04,750 --> 00:06:07,830 How should I think of that? Our belief is that this is a, 113 00:06:07,870 --> 00:06:10,630 this is a generative AI is a participant sport. 114 00:06:10,630 --> 00:06:14,030 You have to jump in and start using it and experience it and 115 00:06:14,510 --> 00:06:17,390 and do some experimentation. So we're encouraging, you know 116 00:06:17,390 --> 00:06:19,830 companies to do that and that's the approach we're taking in our 117 00:06:20,110 --> 00:06:23,260 in our own organization. And but it's also it's very 118 00:06:23,260 --> 00:06:25,740 early with the models that you just, you just highlighted that 119 00:06:25,740 --> 00:06:29,540 with the, you know how young you know the GPT and ChatGPT models 120 00:06:29,540 --> 00:06:32,740 are and a lot of a lot of companies still you know do not 121 00:06:32,740 --> 00:06:34,900 have, you know have not reached GA, you know general 122 00:06:34,900 --> 00:06:37,540 availability status of their models and products. 123 00:06:37,540 --> 00:06:41,180 So this is is the early and evolving Elon Elon's company was 124 00:06:41,180 --> 00:06:43,300 announced recently and there's new companies sprouting up 125 00:06:43,300 --> 00:06:46,740 continuously. And so I think the key for for 126 00:06:46,740 --> 00:06:50,960 companies is is first. Is first you look across your 127 00:06:50,960 --> 00:06:52,920 business and decide where it's applicable. 128 00:06:53,240 --> 00:06:57,240 Second is pick some use cases where you can you we can jump in 129 00:06:57,240 --> 00:06:59,560 and experiment with the technology and manage some of 130 00:06:59,560 --> 00:07:03,340 the complexity and risk. And then third, develop the the, 131 00:07:03,380 --> 00:07:06,460 the foundational capabilities that you need to then scale it 132 00:07:06,460 --> 00:07:08,300 faster. Those capabilities include 133 00:07:08,300 --> 00:07:12,060 technology capabilities like understanding the models, how do 134 00:07:12,060 --> 00:07:14,860 you, you know the prompt engineering, the pretraining and 135 00:07:14,860 --> 00:07:17,620 other things that you might need to do and how to integrate these 136 00:07:17,620 --> 00:07:20,500 models back into your business as well as the business skills 137 00:07:20,500 --> 00:07:23,220 of of understanding how and where you apply it. 138 00:07:23,220 --> 00:07:24,700 How do you develop a business case for it? 139 00:07:24,700 --> 00:07:27,620 How much does it cost to do you know, to apply these models? 140 00:07:28,040 --> 00:07:30,080 And that's really those three steps. 141 00:07:30,080 --> 00:07:32,120 You know, looking, looking across the landscape, 142 00:07:32,120 --> 00:07:35,680 experimenting and laying the foundation are what what we're 143 00:07:35,680 --> 00:07:37,080 helping a lot of companies do today. 144 00:07:37,600 --> 00:07:43,080 Be sure to subscribe to our newsletter, subscribe to our 145 00:07:43,080 --> 00:07:45,600 YouTube channel, check out CXO, talk.com. 146 00:07:45,840 --> 00:07:52,040 Paul, you're describing this kind of open field of innovation 147 00:07:52,040 --> 00:07:56,150 that's going to be happening. But everything around generative 148 00:07:56,150 --> 00:08:00,950 AI right now seems so ambiguous. The technology is changing. 149 00:08:01,310 --> 00:08:07,110 The implications for business are apparently amazing but 150 00:08:07,110 --> 00:08:09,670 unclear. And so how should business 151 00:08:09,670 --> 00:08:12,550 leaders navigate this intense ambiguity? 152 00:08:13,090 --> 00:08:16,330 I think General Bay I is just a new ingredient into the mix. 153 00:08:16,850 --> 00:08:19,010 We've been talking for a while. I've been talking for a while 154 00:08:19,010 --> 00:08:22,010 about the the you know exponential advance in 155 00:08:22,010 --> 00:08:24,530 technology that we're that we're living in. 156 00:08:24,530 --> 00:08:26,450 And so organization you've been talking for a while about 157 00:08:26,450 --> 00:08:29,810 organizations need to develop the ability to innovate it and 158 00:08:29,930 --> 00:08:34,210 and and recognize that adapt technology faster And the the 159 00:08:34,210 --> 00:08:37,970 three key technologies that are really I think will define 160 00:08:37,970 --> 00:08:41,210 companies success in the next several years and decades are 161 00:08:41,210 --> 00:08:43,710 cloud. Artificial intelligence and the 162 00:08:43,710 --> 00:08:47,350 metaverse and those that you know 3 themes that I can talk 163 00:08:47,430 --> 00:08:49,910 more about the you were talking about A I today happy to go to 164 00:08:49,910 --> 00:08:54,550 others other directions as well. And you know, as you look at the 165 00:08:54,550 --> 00:08:57,390 A I piece of it, you know those those things build on each other 166 00:08:57,390 --> 00:08:59,510 to success be successful with A I. 167 00:09:00,120 --> 00:09:02,560 We're finding and companies are finding they need to get to the 168 00:09:02,560 --> 00:09:04,120 cloud. Those that are having advanced 169 00:09:04,120 --> 00:09:07,720 foundation the cloud are better prepared on how to utilize A I 170 00:09:07,800 --> 00:09:10,320 you know most of these models run in, you know in in the cloud 171 00:09:10,800 --> 00:09:13,880 and you need to have your data foundation in place to have your 172 00:09:13,880 --> 00:09:15,800 data. You know have the data to drive 173 00:09:16,160 --> 00:09:19,920 the A I model successfully and a lot of organizations have 174 00:09:19,920 --> 00:09:21,160 struggled with this over the years. 175 00:09:21,160 --> 00:09:24,400 We did a recent survey and only 5 to 10% of the companies. 176 00:09:24,790 --> 00:09:28,030 Really have maturity in terms of how they manage their data and 177 00:09:28,030 --> 00:09:31,630 the and the corresponding A I capabilities that means 90% have 178 00:09:31,630 --> 00:09:34,030 a long way to go. So start with the you know you 179 00:09:34,030 --> 00:09:35,830 need to start with the cloud foundation of what you're 180 00:09:35,830 --> 00:09:37,910 looking at. You need to look at your data 181 00:09:37,910 --> 00:09:39,870 the governance around your data and your metadata and how you 182 00:09:39,870 --> 00:09:43,150 pull that together so that you can you know support you know A 183 00:09:43,150 --> 00:09:46,430 I in the right way and then it's the you know the A I capability 184 00:09:46,750 --> 00:09:48,430 and skills that you build on top of that. 185 00:09:48,710 --> 00:09:50,910 So it's a journey that we're on it but and it's going to 186 00:09:50,910 --> 00:09:52,910 continue. That generative A I is amazing, 187 00:09:52,910 --> 00:09:54,590 but it's not the last big breakthrough, it's. 188 00:09:54,790 --> 00:09:57,030 Not the biggest breakthrough we'll see in technology as this 189 00:09:57,030 --> 00:10:01,030 exponential advance continues. So this is kind of the the, the, 190 00:10:01,030 --> 00:10:02,990 the muscle so to speak. The organizations need to 191 00:10:02,990 --> 00:10:06,150 develop, to continuously anticipate and have the 192 00:10:06,150 --> 00:10:09,150 flexibility in their systems, their architecture and their 193 00:10:09,150 --> 00:10:12,190 business and their business processes and their talent to, 194 00:10:12,230 --> 00:10:14,350 you know, continue to adapt as technology advances. 195 00:10:14,830 --> 00:10:22,030 So from your perspective, AI is essentially another in chain in 196 00:10:22,030 --> 00:10:27,110 a chain of technologies that's not necessarily all that 197 00:10:27,110 --> 00:10:29,430 different from what's come before. 198 00:10:29,870 --> 00:10:33,790 What's different about AI? It is the latest of the chain 199 00:10:33,790 --> 00:10:35,310 and these things all build on each other. 200 00:10:36,270 --> 00:10:39,110 It is this combinatorial effect of the technologies coming 201 00:10:39,110 --> 00:10:40,830 together that really creates the power. 202 00:10:41,510 --> 00:10:44,790 But what's different about AI is it allows us to create more 203 00:10:44,790 --> 00:10:47,590 human like capability. I can communicate. 204 00:10:48,030 --> 00:10:51,710 With with the large language models using natural language, 205 00:10:52,390 --> 00:10:55,230 using voice interaction etcetera, I can get output 206 00:10:55,230 --> 00:10:57,950 that's you know that's easier for me to interpret. 207 00:10:58,190 --> 00:11:01,350 So that's the powerful breakthrough with with the with 208 00:11:01,350 --> 00:11:05,190 generative A I and the more you know what I advocate is the more 209 00:11:05,190 --> 00:11:08,470 human like the technology the more powerful and the more 210 00:11:08,470 --> 00:11:10,670 exciting it is for us. We shouldn't view it as a threat 211 00:11:10,670 --> 00:11:14,270 as technology acquires this capability allows us to really, 212 00:11:14,310 --> 00:11:17,030 you know, to really leverage the technology and and give us you 213 00:11:17,030 --> 00:11:18,950 know, kind of super. Powers is, you know, what we 214 00:11:18,950 --> 00:11:22,470 talked about in our book around giving us new capability. 215 00:11:22,470 --> 00:11:26,190 For example, I can be a customer service Rep and rather than just 216 00:11:26,190 --> 00:11:28,830 what I know in my memory and from my my experience, I can 217 00:11:28,830 --> 00:11:31,390 understand, I can, you know, have out my fingertips. 218 00:11:31,870 --> 00:11:34,710 Every aspect of every technical manual on the product that I'm 219 00:11:34,710 --> 00:11:37,630 answering questions on brought me a brought to the forefront 220 00:11:37,630 --> 00:11:39,830 prioritize that I can answer the customer's question the right 221 00:11:39,830 --> 00:11:41,830 way. This is the type of power you 222 00:11:41,830 --> 00:11:44,670 know, that's that the technology's given us and, you 223 00:11:44,670 --> 00:11:47,460 know, just to. You know to go at that a little 224 00:11:47,460 --> 00:11:51,500 further, when we look at the real impact of a I while you 225 00:11:51,500 --> 00:11:53,980 know cloud probably changed technology a lot and how we 226 00:11:53,980 --> 00:11:55,740 built technology and supported technology. 227 00:11:55,780 --> 00:11:59,380 A I changing work and the way we work because of this capability. 228 00:11:59,660 --> 00:12:03,100 And one of this, the research studies we did recently showed 229 00:12:03,100 --> 00:12:07,900 that 40% of working hours across companies globally. 230 00:12:08,450 --> 00:12:11,050 Are impacted by generative AI 40%. 231 00:12:11,250 --> 00:12:13,930 That doesn't mean 40% of jobs go away, far from it. 232 00:12:14,130 --> 00:12:16,970 We actually see it enhancing jobs and enhancing productivity 233 00:12:16,970 --> 00:12:19,770 capabilities that people have in in many ways and you'll have to 234 00:12:19,770 --> 00:12:23,290 go into that in more detail. Q Harrison mentioned that your 235 00:12:23,290 --> 00:12:28,010 book was called Human Plus Machine, and we have a really 236 00:12:28,010 --> 00:12:32,130 interesting question from LinkedIn, and this is from 237 00:12:32,330 --> 00:12:35,210 Melena Z. And, she says, how would you 238 00:12:35,210 --> 00:12:40,530 describe the significance of incorporating human values into 239 00:12:40,530 --> 00:12:44,210 the development of generative AI technology? 240 00:12:44,610 --> 00:12:48,470 It's incredibly important. It's if if you don't have in 241 00:12:48,470 --> 00:12:52,470 your organization a really strong responsible A I program 242 00:12:52,750 --> 00:12:54,470 you're simply being irresponsible. 243 00:12:54,750 --> 00:12:58,630 And at the core of responsible A I is you know counting for human 244 00:12:58,630 --> 00:13:01,350 values and in the way that you do it. 245 00:13:01,470 --> 00:13:06,510 Responsible A I in our view is about it's about things like the 246 00:13:06,630 --> 00:13:10,910 you know accuracy and and coming up with the right answers, 247 00:13:10,910 --> 00:13:15,230 avoiding the hallucination. It's about the ethical issues 248 00:13:15,230 --> 00:13:16,670 that cut that you need to think about. 249 00:13:16,830 --> 00:13:19,190 In terms of how you're applying the A I, it's about bias and 250 00:13:19,190 --> 00:13:22,590 ensuring you that there's fair you know fair outcomes and fair 251 00:13:22,590 --> 00:13:26,630 use of the technology and and you know where in in certain 252 00:13:26,630 --> 00:13:29,110 cases the transparency and explainability that you need 253 00:13:29,430 --> 00:13:33,990 around the technology and we're we encourage every organization 254 00:13:33,990 --> 00:13:37,030 using a I to do is is really especially with the advance of 255 00:13:37,030 --> 00:13:38,990 generative A I we've been talking about this for six years 256 00:13:39,190 --> 00:13:41,750 but especially with generative A I, you need a responsible A I 257 00:13:41,750 --> 00:13:45,150 program in place and if you can inventory every use of a I in 258 00:13:45,150 --> 00:13:47,680 your company. And understand the risk of it 259 00:13:47,840 --> 00:13:50,280 and know how you're mitigating those risks, then you're simply 260 00:13:50,280 --> 00:13:53,680 going to get yourself in trouble with the with improper uses of a 261 00:13:53,680 --> 00:13:57,280 I That's that's what the way we think about responsible a I it's 262 00:13:57,280 --> 00:14:00,480 not just, you know, mushy values and principles. 263 00:14:00,760 --> 00:14:04,120 It's execution, operations and compliance in terms of how 264 00:14:04,120 --> 00:14:07,510 you're applying the technology. I mean, Paul, it's a great 265 00:14:07,510 --> 00:14:11,150 point, but the question I have is like the theoretical version 266 00:14:11,150 --> 00:14:13,950 of that and the actual application of that often times 267 00:14:13,950 --> 00:14:17,550 look entirely different. For example, if I'm in a company 268 00:14:17,550 --> 00:14:20,550 and let's say I'm experimenting with generative AI and it's just 269 00:14:20,550 --> 00:14:23,590 in our R&D department. And then we quickly realized 270 00:14:23,590 --> 00:14:25,390 that this could actually have some scale. 271 00:14:26,390 --> 00:14:30,230 We just apply it to a whole another sector of our company, 272 00:14:30,230 --> 00:14:31,870 or maybe we apply it to the whole company. 273 00:14:33,050 --> 00:14:36,450 At what point do I actually stop and and say okay, there's a 274 00:14:36,450 --> 00:14:39,570 legal component here and often times when we point to that 275 00:14:39,570 --> 00:14:43,210 direction, I mean we're that's the big debate in a I today. 276 00:14:43,210 --> 00:14:46,090 I mean even at the congressional level is you know, what do we 277 00:14:46,090 --> 00:14:47,570 do? How do we regulate this stuff? 278 00:14:47,770 --> 00:14:50,890 If I stop now, aren't I hindering my innovation? 279 00:14:50,890 --> 00:14:53,850 And if I'm in charge of innovation and acceleration of 280 00:14:53,850 --> 00:14:57,850 technological development within the company, what like I'm 281 00:14:57,850 --> 00:15:00,250 caught in the catch 22, if you, if you understand what I'm 282 00:15:00,250 --> 00:15:02,510 saying? I am not one of those that 283 00:15:02,510 --> 00:15:06,270 supports stopping and banning or pausing on the technology. 284 00:15:06,470 --> 00:15:09,870 I think it's about putting in place the right framework, in 285 00:15:09,870 --> 00:15:13,110 the right guidelines that you know what you're doing and can 286 00:15:13,110 --> 00:15:15,270 evaluate the risk of it. I would say not just at the end, 287 00:15:15,270 --> 00:15:17,630 but every step along the way and before you even get started, you 288 00:15:17,630 --> 00:15:19,110 should do an assessment. Of the risk. 289 00:15:19,630 --> 00:15:22,030 You know, there's a lot of guidelines and and ways you can 290 00:15:22,030 --> 00:15:23,910 do that. The EUEU is. 291 00:15:24,740 --> 00:15:27,700 Is going through the stages of approval on the A I act, they 292 00:15:27,700 --> 00:15:30,580 identify high risk you know different risk categories of a I 293 00:15:31,180 --> 00:15:34,220 do you know does your team understand those and are you 294 00:15:34,260 --> 00:15:37,220 assessing for any application of a I whether what risk category 295 00:15:37,220 --> 00:15:40,100 you're fitting into and then how you you know mitigate that or 296 00:15:40,100 --> 00:15:42,020 deal with that to make sure you're you're handling that. 297 00:15:42,020 --> 00:15:44,580 And that's just one example of respective EU there's also the 298 00:15:44,580 --> 00:15:46,980 White House guidelines, there's NIST and and other things that 299 00:15:46,980 --> 00:15:49,220 are that are out there and they'll and they'll be more 300 00:15:49,220 --> 00:15:51,660 coming because of the interest in in kind of setting some 301 00:15:51,660 --> 00:15:54,580 guardrails around this which I think is a good thing but. 302 00:15:54,740 --> 00:15:56,980 But I think, I think the teams need to be trained and 303 00:15:56,980 --> 00:16:00,500 organizations need to have tools in place so that you see you are 304 00:16:00,500 --> 00:16:02,900 assessing you know the use of a I and understand again 305 00:16:02,900 --> 00:16:05,300 understand that the risk of it and make decisions accordingly 306 00:16:05,300 --> 00:16:06,660 there. There's things we won't do and 307 00:16:06,660 --> 00:16:10,900 things we've decided not to do. You know applications that we've 308 00:16:11,500 --> 00:16:14,460 not pursued because of you know the the risk profile of more we 309 00:16:14,540 --> 00:16:16,580 we did. We felt it was not aligned with 310 00:16:16,660 --> 00:16:19,700 the the right values and that's an important consideration to 311 00:16:19,700 --> 00:16:22,990 build into your process. It can't just be after the fact. 312 00:16:22,990 --> 00:16:25,590 It's got to be, you know, as you're considering use cases and 313 00:16:25,750 --> 00:16:29,870 starting out. AI It brings out either the best 314 00:16:29,870 --> 00:16:34,070 in people or the worst in people, and the latter component 315 00:16:34,070 --> 00:16:35,550 when it brings out the worst in people. 316 00:16:36,160 --> 00:16:39,040 Traditionally what I'm seeing is people will try to hinder the 317 00:16:39,160 --> 00:16:43,560 Ai's abilities or slow it down in fear of losing their job or 318 00:16:44,040 --> 00:16:46,440 seeing other calamities ensue within their industry. 319 00:16:46,640 --> 00:16:49,880 One of the questions I have for you is how do we get better at 320 00:16:49,880 --> 00:16:52,970 communicating AI? Technology is neutral. 321 00:16:52,970 --> 00:16:54,770 Like technology isn't good or bad. 322 00:16:54,890 --> 00:17:00,090 There's yeah in in AI fit it was generative. 323 00:17:00,090 --> 00:17:02,170 AI fits in fits into that description. 324 00:17:02,170 --> 00:17:05,250 Gender value isn't good or bad. It's exactly as you said, Q. 325 00:17:05,250 --> 00:17:08,609 It's how you use it and it can be used for bad purposes. 326 00:17:08,609 --> 00:17:11,210 It could be used to spread misinformation at scale, deep 327 00:17:11,210 --> 00:17:14,849 fakes and all sorts of things. But that's people using the 328 00:17:14,849 --> 00:17:18,369 technology that I think people I think that's what some of the 329 00:17:18,369 --> 00:17:21,970 communication we need to do around generative A I is that 330 00:17:21,970 --> 00:17:25,089 the thing we really need to be looking out for and preventing 331 00:17:25,290 --> 00:17:28,369 is bad uses of a I and people using a I in bad ways. 332 00:17:28,369 --> 00:17:30,850 We need to educate, you know, the broader, you know, the 333 00:17:30,850 --> 00:17:33,690 general population on what that means that they they can 334 00:17:33,730 --> 00:17:36,250 recognize and understand if something you know has been 335 00:17:36,730 --> 00:17:38,970 appropriate propagated and generated, you know. 336 00:17:39,400 --> 00:17:42,000 You know, artificially at scale using gender of a I for some 337 00:17:42,720 --> 00:17:44,320 illicit purpose whatever it might be. 338 00:17:44,760 --> 00:17:48,040 So I I think that's like there there is a broad education that 339 00:17:48,040 --> 00:17:50,360 that needs to happen that we're doing a lot of work on that 340 00:17:50,360 --> 00:17:53,040 we're working with a lot of different organizations on that 341 00:17:53,040 --> 00:17:57,880 governments and other bodies to you know to look at how we can 342 00:17:58,120 --> 00:18:00,920 better. Educate, you know, the the 343 00:18:00,920 --> 00:18:03,000 general population, as well as business leaders and 344 00:18:03,000 --> 00:18:04,600 technologists and decision makers. 345 00:18:05,080 --> 00:18:06,920 Around using the technology in the right way. 346 00:18:07,200 --> 00:18:11,720 So I think that's an ongoing effort that we'll all need to 347 00:18:11,840 --> 00:18:14,600 work together on. We have a bunch of questions 348 00:18:14,600 --> 00:18:18,360 that are stacking up on LinkedIn and Twitter. 349 00:18:18,360 --> 00:18:22,680 And I have to say you guys in the audience, you are so 350 00:18:22,680 --> 00:18:26,200 intelligent, so smart and sophisticated and your questions 351 00:18:26,200 --> 00:18:31,210 are absolutely great. And our next question comes from 352 00:18:31,210 --> 00:18:33,690 Florin Rotar. He is the Chief Technology 353 00:18:33,690 --> 00:18:38,370 Officer at Avanade. And I have to say that I did a 354 00:18:38,530 --> 00:18:43,050 video with Florin years and years and years and years ago in 355 00:18:43,050 --> 00:18:45,530 Seattle. So Florin, it's great to see you 356 00:18:45,690 --> 00:18:48,410 pop up. And here's here's Florin's 357 00:18:48,410 --> 00:18:51,330 question and I think it gets right to the heart of some of 358 00:18:51,330 --> 00:18:54,850 the key issues. And he says, how will Generative 359 00:18:54,970 --> 00:18:59,720 A I change the future of work? Can it also play a role to 360 00:18:59,720 --> 00:19:03,680 enable people to realize their full potential to thrive and to 361 00:19:03,680 --> 00:19:07,640 grow, not just to drive productivity? 362 00:19:08,560 --> 00:19:13,480 Will it blur the lines between white color and blue color? 363 00:19:14,240 --> 00:19:17,360 And I'll just add to that. To me, this question is also 364 00:19:17,360 --> 00:19:20,960 getting to the point that Q Harrison just raised, which is 365 00:19:20,960 --> 00:19:24,800 generative. A I brings out the the best and 366 00:19:24,800 --> 00:19:29,410 brings out the worst in people. We talked in human plus machine 367 00:19:29,410 --> 00:19:33,210 about the idea of no collar jobs and exactly what the foreign 368 00:19:33,210 --> 00:19:37,010 highlights of eliminating this distinction between you know 369 00:19:37,090 --> 00:19:38,930 kind of blue collar, white white collar. 370 00:19:38,930 --> 00:19:42,050 As you look at it. I mean think about it a hands on 371 00:19:42,050 --> 00:19:45,730 service technician, think about a plumber or an electrician that 372 00:19:45,730 --> 00:19:49,090 now has access to large language models that that give them 373 00:19:49,090 --> 00:19:51,530 tremendous amounts of additional information and potential. 374 00:19:51,530 --> 00:19:54,210 It can give them tools to run their business more effectively. 375 00:19:54,210 --> 00:19:56,730 Maybe they can be a a service provider to others in their 376 00:19:56,730 --> 00:19:59,170 profession rather than. Than just you know, being rather 377 00:19:59,170 --> 00:20:02,370 than just being you know the specialist at the the physical 378 00:20:02,370 --> 00:20:04,970 trade that they have. I think that's the blurring of 379 00:20:05,170 --> 00:20:07,210 capability that the the A I allows. 380 00:20:07,490 --> 00:20:10,690 And think about it, a small business that now or at any part 381 00:20:10,690 --> 00:20:13,730 of a larger business that wants to go international overnight, 382 00:20:13,730 --> 00:20:16,650 they can start communicating in dozens of languages. 383 00:20:17,340 --> 00:20:19,980 You know seamlessly and expanding their business, you 384 00:20:19,980 --> 00:20:22,820 know it's the Super powers that are enabled that give people 385 00:20:22,820 --> 00:20:25,940 more capability and that leads to a lot of you know, new 386 00:20:25,940 --> 00:20:27,660 entrepreneurial activity and ideas. 387 00:20:27,940 --> 00:20:30,540 I mean think of what GoDaddy did to the Internet and creating a 388 00:20:30,540 --> 00:20:33,740 generation of entrepreneurs a lot of different ways or eBay 389 00:20:33,740 --> 00:20:36,580 market marketplaces and such. We're going to see that you know 390 00:20:36,620 --> 00:20:39,860 to the next, you know to the next exponential multiple with 391 00:20:39,860 --> 00:20:43,020 generative AI creating all these new possibilities of what people 392 00:20:43,300 --> 00:20:45,700 can do. So that's the kind of what what 393 00:20:45,700 --> 00:20:50,260 we see happening there more specific around it, we we see 394 00:20:50,460 --> 00:20:53,460 you know the new opportunities for jobs and white generative A 395 00:20:53,460 --> 00:20:55,780 I impacts that falling into 5 categories. 396 00:20:56,340 --> 00:20:59,580 The 1st is advising and this is kind of you know advisors or 397 00:20:59,580 --> 00:21:02,570 assistants or copilots to help people do their jobs more 398 00:21:02,570 --> 00:21:04,730 effectively. A large for example large 399 00:21:04,730 --> 00:21:07,410 European service organization that we're working with. 400 00:21:07,770 --> 00:21:10,370 We're using generative A I in the customer service 401 00:21:10,370 --> 00:21:13,130 organization to allow them to answer questions a lot more 402 00:21:13,130 --> 00:21:15,890 accuracy and quality because they can as I mentioned earlier, 403 00:21:15,890 --> 00:21:18,370 they can, they can pull tremendous amounts of technical 404 00:21:18,370 --> 00:21:20,650 information together to answer customers questions better, 405 00:21:20,650 --> 00:21:23,450 faster with higher quality And they can cross sell more 406 00:21:23,450 --> 00:21:26,890 effectively because they get the ideas and prompts and support On 407 00:21:26,890 --> 00:21:29,410 on how to cross sell that's you know, advising. 408 00:21:29,690 --> 00:21:32,250 Creating is another whole count is a second whole category. 409 00:21:32,250 --> 00:21:36,970 Another category, A good example here is work we're doing with in 410 00:21:36,970 --> 00:21:41,330 a in the Pharmaceutical industry where we're able to in the drug 411 00:21:41,330 --> 00:21:44,720 discovery process and clinical trials process create some of 412 00:21:44,720 --> 00:21:47,200 the regulatory and compliance documents they need to create. 413 00:21:47,480 --> 00:21:50,280 So that then gets reviewed at the final stage by humans in the 414 00:21:50,280 --> 00:21:53,840 loop avoiding all the road work and you know that a person would 415 00:21:53,840 --> 00:21:56,240 normally do and allowing them to apply their judgment and 416 00:21:56,240 --> 00:21:59,520 expertise in the final product. That's creative in addition to 417 00:21:59,880 --> 00:22:02,520 you know applying it in marketing and other areas that I 418 00:22:02,520 --> 00:22:04,400 can talk about which is super interesting right now. 419 00:22:04,840 --> 00:22:06,400 So that's the creating side of it. 420 00:22:06,400 --> 00:22:09,080 There's automating where you can use Gender VI to automate some 421 00:22:09,080 --> 00:22:11,640 of the transaction processing. An example here is a 422 00:22:11,640 --> 00:22:14,440 multinational bank. We're using generative A I in 423 00:22:14,440 --> 00:22:18,360 their in their back office processing to correlate read and 424 00:22:18,360 --> 00:22:20,920 correlate 10s of thousands of emails that come in with 425 00:22:20,920 --> 00:22:23,560 transaction activity. Normally people need to sort 426 00:22:23,560 --> 00:22:26,840 through all this to reconcile and do their you know their 427 00:22:26,920 --> 00:22:29,440 their their post trade processing more effectively. 428 00:22:29,800 --> 00:22:32,440 Again, you can do this with other technology, you can do it 429 00:22:32,440 --> 00:22:34,840 with gender A I and you can make people's jobs that more 430 00:22:34,840 --> 00:22:36,760 productive and effective. And you know take out some of 431 00:22:36,760 --> 00:22:40,800 the the drudgery, the 4th category that is protecting, 432 00:22:41,220 --> 00:22:42,460 which I think is super interesting. 433 00:22:42,460 --> 00:22:45,900 An example here is we're working with a large energy company on a 434 00:22:45,900 --> 00:22:49,340 safety application so that workers in real time can get all 435 00:22:49,340 --> 00:22:52,420 the information on what's happening real time conditions 436 00:22:52,420 --> 00:22:56,220 weather conditions and other things in a complex say refinery 437 00:22:56,940 --> 00:22:59,740 and and then combine that with the all the information they 438 00:22:59,740 --> 00:23:02,940 need to know from safety procedures and manuals and and 439 00:23:02,980 --> 00:23:06,140 regulation and such that they can they can operate in a more 440 00:23:06,140 --> 00:23:09,540 safe manner in real time. You know again couldn't do this 441 00:23:09,660 --> 00:23:11,980 all put all this together for generative A, I and then the 442 00:23:11,980 --> 00:23:15,460 final use case we're seeing a lot is in technology itself 443 00:23:15,460 --> 00:23:17,580 using a I and software development and technology 444 00:23:17,580 --> 00:23:19,180 development. I'm sure we'll find more 445 00:23:19,180 --> 00:23:21,740 examples as we go. Those are five, they're kind of 446 00:23:21,740 --> 00:23:24,620 standing out right now just to you know drill into you know 447 00:23:24,620 --> 00:23:26,380 some of the ways that it's transforming work in the 448 00:23:26,420 --> 00:23:29,140 response before into question. We've got another question 449 00:23:29,140 --> 00:23:33,140 coming in from Twitter from Chris Peterson, and the question 450 00:23:33,140 --> 00:23:36,380 is, one of the opportunities mentioned in Human Plus Machine 451 00:23:36,700 --> 00:23:40,020 was the AI explainer role. Is that even possible for 452 00:23:40,020 --> 00:23:43,500 something as complex as GPT 4 with billions of parameters and 453 00:23:43,500 --> 00:23:46,730 almost unlimited training data in some industries and some 454 00:23:46,730 --> 00:23:49,050 problems If you can't explain it, you can't do it. 455 00:23:49,250 --> 00:23:51,490 That's part of that screening that I talked about earlier with 456 00:23:51,530 --> 00:23:53,810 responsible A I. If you have a kind of a 457 00:23:53,810 --> 00:23:57,210 regulatory or ethical or business need to explain exactly 458 00:23:57,210 --> 00:23:58,810 how things something's happening. 459 00:23:59,010 --> 00:24:01,330 You need to use the right type of approach where you can do 460 00:24:01,330 --> 00:24:03,850 that and you can't do that with some to your point with some of 461 00:24:03,850 --> 00:24:05,450 the the models that that are there. 462 00:24:05,730 --> 00:24:07,890 But there's a lot of advance happening to explain ability. 463 00:24:07,890 --> 00:24:10,250 There's ways to query the models to understand, you know, how 464 00:24:10,250 --> 00:24:15,960 they're processing. There's areas like gender Gans, 465 00:24:15,960 --> 00:24:19,160 gender adversarial networks we can use in different ways to get 466 00:24:19,160 --> 00:24:21,680 some insight into how models are working and such. 467 00:24:21,960 --> 00:24:24,520 There's a lot of different advances there and there are, 468 00:24:24,920 --> 00:24:28,560 you know, there's new fields in addition, new fields like prompt 469 00:24:28,560 --> 00:24:31,240 engineering that are cropping up because of gender of a I we're 470 00:24:31,240 --> 00:24:33,920 also seeing, you know, demands the market for explainability 471 00:24:34,320 --> 00:24:37,080 engineers or explainability specialists who can bring that 472 00:24:37,080 --> 00:24:41,880 understanding in to help them, to help understand those kind of 473 00:24:41,880 --> 00:24:43,800 conditions. And the other thing that's 474 00:24:43,800 --> 00:24:47,360 sometimes important is that even if in some applications you 475 00:24:47,360 --> 00:24:49,720 don't necessarily need to explain exactly how you got the 476 00:24:49,760 --> 00:24:53,020 answer, you need to provide the transparency of what information 477 00:24:53,020 --> 00:24:55,780 you're using, what data are you using, and and the process 478 00:24:55,780 --> 00:24:57,700 itself for. So you need to differentiate 479 00:24:57,700 --> 00:25:00,620 where do you really need to explain exactly all the math you 480 00:25:00,620 --> 00:25:02,100 did and how you did it, so to speak? 481 00:25:02,500 --> 00:25:05,020 And where do you just need to provide transparency to how 482 00:25:05,020 --> 00:25:07,140 you're doing it and show that you're using information such in 483 00:25:07,140 --> 00:25:10,260 the right way and distinguishing that can you know can help 484 00:25:10,260 --> 00:25:12,500 organizations unlock some of the potential to. 485 00:25:13,020 --> 00:25:16,980 And we have another question from Wayne Anderson. 486 00:25:17,430 --> 00:25:22,310 And you can see we love taking questions from the audience and 487 00:25:22,310 --> 00:25:24,910 and again the audience is amazing. 488 00:25:25,550 --> 00:25:31,470 So, so this is from Wayne Anderson on Twitter and Wayne 489 00:25:31,470 --> 00:25:34,350 also has a question coming up on LinkedIn. 490 00:25:34,350 --> 00:25:38,710 So he's like sort of a multi tenanted multifaceted. 491 00:25:38,710 --> 00:25:44,150 It's social media happening here and and Wayne says what is the 492 00:25:44,150 --> 00:25:46,670 litmus test? Is there one? 493 00:25:47,780 --> 00:25:52,380 A question set of questions that you use to quickly evaluate A 494 00:25:52,380 --> 00:25:56,300 client's place on the operational maturity journey. 495 00:25:56,300 --> 00:26:01,300 For AI and ML, we have a maturity framework We use to 496 00:26:01,300 --> 00:26:06,340 assess for ourselves as well As for our clients that it there's 497 00:26:06,340 --> 00:26:11,490 steps of maturity you know that that you go through in in 498 00:26:11,490 --> 00:26:13,970 assessing it. There's there's assessing you 499 00:26:13,970 --> 00:26:17,730 know talent and where you are with the with the talent and the 500 00:26:17,810 --> 00:26:19,450 expertise that you have in the organization. 501 00:26:19,450 --> 00:26:21,650 That's about the technology talent as well as you know the 502 00:26:21,730 --> 00:26:24,210 skills you have in the business and the kind of trading programs 503 00:26:24,210 --> 00:26:27,210 that you have around that. There's assessing the data 504 00:26:27,210 --> 00:26:31,130 readiness that for it you know in terms of as we talked about 505 00:26:31,130 --> 00:26:35,210 earlier, your data maturity and the maturity of platforms, data 506 00:26:35,210 --> 00:26:37,290 platforms to support what you need to do. 507 00:26:37,810 --> 00:26:43,650 There's that. There's then the maturity of how 508 00:26:43,650 --> 00:26:46,210 you what do you, how you need to use the models and your 509 00:26:46,250 --> 00:26:49,250 sophistication around that. And that depends on the strategy 510 00:26:49,250 --> 00:26:53,330 that you have. Is your strategy to use, use you 511 00:26:53,330 --> 00:26:57,120 know, proprietary, you know, pre trade public, you know available 512 00:26:57,120 --> 00:26:59,200 models. There's your strategy to do some 513 00:26:59,200 --> 00:27:01,400 of your own pre training or customization using your own 514 00:27:01,400 --> 00:27:04,520 data that requires far different operational skills. 515 00:27:04,520 --> 00:27:07,480 And therefore you you need to evaluate where you are on that, 516 00:27:07,840 --> 00:27:10,480 on that spectrum. And then there's the, the 517 00:27:10,480 --> 00:27:13,400 operate the operational skills around it. 518 00:27:13,400 --> 00:27:16,760 So how do you, how do you put the A I in place and how do you 519 00:27:16,760 --> 00:27:19,860 moderate on an ongoing basis for the right outcomes? 520 00:27:19,900 --> 00:27:22,260 And then finally, the responsible AI dimension of it. 521 00:27:22,260 --> 00:27:23,340 So those are kind of the dimensions. 522 00:27:23,340 --> 00:27:27,620 There's more underneath that, but there's a, there's a process 523 00:27:27,620 --> 00:27:29,580 that we used to go to, to go through it. 524 00:27:29,580 --> 00:27:31,140 I think that's every organization. 525 00:27:31,140 --> 00:27:33,540 Having an understanding of that and having a way to evaluate 526 00:27:33,820 --> 00:27:36,580 their maturity is important to to know how you're making 527 00:27:36,580 --> 00:27:39,260 progress. Wayne actually did ask another 528 00:27:39,260 --> 00:27:41,780 question on LinkedIn. I'm looking at it right here and 529 00:27:41,780 --> 00:27:44,940 he talks about the security and the risk of AI is not something 530 00:27:44,940 --> 00:27:46,900 that is entirely a technical solution. 531 00:27:47,330 --> 00:27:50,450 A lot of it is in the humans and the innovation back slash 532 00:27:50,450 --> 00:27:54,370 development processes and what formal steps do you need to be 533 00:27:54,650 --> 00:27:59,850 in order for that innovation to provide the kind of guide rails 534 00:28:00,090 --> 00:28:03,010 and talking points on the future of machine learning projects. 535 00:28:03,290 --> 00:28:08,210 So the way I the way I interpret that is, you know we've got a 536 00:28:08,210 --> 00:28:12,410 lot of groups working together. How do you, how do you make sure 537 00:28:12,410 --> 00:28:15,170 they're all working and and their energies are going the 538 00:28:15,170 --> 00:28:17,110 right way? Right direction. 539 00:28:17,390 --> 00:28:20,710 We think the right approach to use is a center of excellence 540 00:28:20,710 --> 00:28:23,110 kind of approach given that the state of the technology where 541 00:28:23,110 --> 00:28:25,630 you create a center of excellence, you know you have to 542 00:28:25,710 --> 00:28:28,150 you know centralize in your organization that has those 543 00:28:28,150 --> 00:28:30,230 capabilities in it. That's what we've done for 544 00:28:30,230 --> 00:28:32,310 ourselves and we're helping a lot of our clients to, in fact 545 00:28:32,310 --> 00:28:34,950 we have something called the Coe in a box that we're using to 546 00:28:34,950 --> 00:28:37,350 help clients set up these kind of capabilities. 547 00:28:37,350 --> 00:28:40,150 It requires the technology capability, the business 548 00:28:40,150 --> 00:28:44,350 capability, the legal, you know, the legal teams and capability, 549 00:28:44,350 --> 00:28:50,390 legal and commercial. And and you know talent, you 550 00:28:50,390 --> 00:28:53,910 know, kind of you know talent HR kind of capability around it. 551 00:28:54,150 --> 00:28:55,710 So you need to. Bring all that, all that 552 00:28:55,710 --> 00:28:58,350 together. So the center of excellence, 553 00:28:58,550 --> 00:29:01,590 where you can have that capability assembled, you have 554 00:29:01,590 --> 00:29:03,230 representatives from all those different groups in your 555 00:29:03,230 --> 00:29:05,990 organization is important. You can federate some of the 556 00:29:05,990 --> 00:29:07,950 experimentation. Then it's really important to 557 00:29:07,950 --> 00:29:09,990 bring it together. Security is a great angle. 558 00:29:10,030 --> 00:29:12,190 I don't know if that was the primary thrust to that question. 559 00:29:12,960 --> 00:29:15,720 But there's a lot of implications on security from 560 00:29:15,720 --> 00:29:19,000 generative A I both in terms of new security challenges as well 561 00:29:19,000 --> 00:29:22,320 as consideration about data privacy, ground grounding of 562 00:29:22,320 --> 00:29:25,280 models, use of sovereign data depending on the jurisdictions 563 00:29:25,280 --> 00:29:28,040 you're operating and that become really critical considerations 564 00:29:28,040 --> 00:29:30,920 for companies. So having this built into you 565 00:29:30,920 --> 00:29:33,080 know kind of a center of excellence that you know that 566 00:29:33,080 --> 00:29:35,200 you're channeling this in the right way in companies, I think 567 00:29:35,200 --> 00:29:38,720 it's critical for the stage of development that we're that 568 00:29:38,720 --> 00:29:41,850 we're at right now. So Paul, giving your purview and 569 00:29:41,850 --> 00:29:45,330 and some of your thesis around the future, one of the things 570 00:29:45,330 --> 00:29:48,450 that I I'm wondering in your your realm is when I look at 571 00:29:48,450 --> 00:29:51,930 technologies like the cloud and enterprise corporation is 572 00:29:51,930 --> 00:29:55,290 probably best suited for that realization today, right. 573 00:29:55,290 --> 00:29:58,290 Like the personal cloud computing it exists and you know 574 00:29:58,290 --> 00:30:00,810 I think the strongest use case of that is probably video games 575 00:30:00,810 --> 00:30:03,410 today. But beyond that, it doesn't make 576 00:30:03,490 --> 00:30:05,730 that much sense for an individual person, or even a 577 00:30:05,730 --> 00:30:09,690 small startup to to endeavor on a very complex. 578 00:30:09,890 --> 00:30:12,410 Cloud implementation. However, I think that might 579 00:30:12,410 --> 00:30:15,570 differ given your some of your comments you just specified when 580 00:30:15,570 --> 00:30:18,370 it comes to AI. On the AI front, one of the 581 00:30:18,370 --> 00:30:22,620 things that we're seeing is. Corporations that have a lot of 582 00:30:22,620 --> 00:30:25,620 technical debt or have a lot of data that hasn't been digitized 583 00:30:25,620 --> 00:30:29,580 or have you know very complex teams and org charts, they're 584 00:30:29,580 --> 00:30:31,980 not well suited because it's going to take them some time to 585 00:30:31,980 --> 00:30:35,180 get all these things in progress in place. 586 00:30:35,340 --> 00:30:37,500 Now on the flip side, they have the most data so they'll 587 00:30:37,500 --> 00:30:40,380 probably have some of the the stronger AI models. 588 00:30:40,580 --> 00:30:43,700 But to what I to the to make the question that I have here is 589 00:30:43,700 --> 00:30:46,340 like would it make more sense for a startup or even an 590 00:30:46,380 --> 00:30:49,420 organization to think about, you know, creating an internal 591 00:30:49,420 --> 00:30:51,620 startup and then going? And after it, I mean, that's 592 00:30:51,620 --> 00:30:55,260 what Google did with DeepMind. And I mean we just saw some of 593 00:30:55,260 --> 00:30:59,860 the new news related to DeepMind where they're bringing in demis 594 00:30:59,900 --> 00:31:03,820 and to lead their actual AI practices at Google. 595 00:31:04,060 --> 00:31:07,100 And even if there's countless examples where this is also true 596 00:31:07,100 --> 00:31:10,300 in the AI industry, is that the right approach or do you think 597 00:31:10,300 --> 00:31:13,540 that that ship has sailed to a long time ago? 598 00:31:14,060 --> 00:31:17,260 For some companies, there's an example, a media organization 599 00:31:17,260 --> 00:31:19,980 we're working with that sees an opportunity. 600 00:31:20,410 --> 00:31:22,530 To really create a whole new part of their business using 601 00:31:22,530 --> 00:31:27,690 gender of a I, they can use gender of a I to create a way to 602 00:31:27,690 --> 00:31:30,090 generate coverage for things they couldn't cover before. 603 00:31:30,090 --> 00:31:34,250 I can't get too too specific about it and and in that case 604 00:31:34,250 --> 00:31:35,810 that's that's maybe more of a startup thing. 605 00:31:35,810 --> 00:31:38,410 You actually are using gender of a I to branch out in a new 606 00:31:38,410 --> 00:31:41,130 direction. But we think a lot of the a lot 607 00:31:41,130 --> 00:31:43,650 of the gender of a I potential is going to be in the changing 608 00:31:43,650 --> 00:31:45,490 the core of how you work as a company. 609 00:31:45,490 --> 00:31:47,410 It's going to transform the way work is done. 610 00:31:47,810 --> 00:31:50,210 That phrase UW is reimagining work. 611 00:31:50,410 --> 00:31:52,690 That's what this is about, which means I think you do need to to 612 00:31:53,090 --> 00:31:55,930 have a lot of capability at the heart of your organization 613 00:31:56,570 --> 00:31:59,370 looking at how you, how you do this, how you do and drive the 614 00:31:59,370 --> 00:32:03,290 transformation. So, so I think it could be a mix 615 00:32:03,290 --> 00:32:04,450 for different types of use cases. 616 00:32:04,490 --> 00:32:08,050 You know, some a company may spin out or have a little, you 617 00:32:08,050 --> 00:32:11,410 know, separate projects you know, to pursue some initiatives 618 00:32:11,410 --> 00:32:12,730 they're doing. But I think this gets to the 619 00:32:12,730 --> 00:32:15,690 core of how companies are operating, which is why, you 620 00:32:15,690 --> 00:32:17,530 know, companies need to embrace it broadly. 621 00:32:17,810 --> 00:32:19,570 But another point that you're mentioning is. 622 00:32:20,400 --> 00:32:24,000 I do think this that the gender they offers a lot of potential 623 00:32:24,000 --> 00:32:26,160 for, for new startups and small companies because they can 624 00:32:26,160 --> 00:32:29,760 access tremendous capability to build new businesses in addition 625 00:32:29,760 --> 00:32:31,240 to the power it gives big businesses. 626 00:32:31,240 --> 00:32:34,560 So I heard people ask does this, you know, are the big only, you 627 00:32:34,560 --> 00:32:37,200 know big companies only going to get stronger with this or are 628 00:32:37,200 --> 00:32:39,440 the did the new startups, you know new companies going to win 629 00:32:39,440 --> 00:32:40,840 out? I think it's really a mix here 630 00:32:41,360 --> 00:32:43,280 that we'll see going forward because of the power of the 631 00:32:43,280 --> 00:32:45,320 models, the power for new organizations to leverage them 632 00:32:45,320 --> 00:32:48,280 as well as, you know, the power that larger organizations have 633 00:32:48,280 --> 00:32:51,180 to move faster. Well, let's shift gears a little 634 00:32:51,180 --> 00:32:57,620 bit and talk about investment. Technology investment A I is 635 00:32:57,620 --> 00:33:02,100 changing so rapidly. The capabilities are changing, 636 00:33:02,100 --> 00:33:07,140 the models are changing. The implications for the 637 00:33:07,180 --> 00:33:12,060 enterprise and for society at large remain very unclear. 638 00:33:12,740 --> 00:33:17,620 Given this ambiguity, how should how do you recommend? 639 00:33:18,020 --> 00:33:22,780 That organizations should be investing and I will mention 640 00:33:22,780 --> 00:33:27,220 that Accenture recently announced a $3 billion 641 00:33:27,340 --> 00:33:30,020 investment in this. So obviously it's something that 642 00:33:30,060 --> 00:33:31,260 you're giving a lot of thought to. 643 00:33:31,780 --> 00:33:34,660 As you said, we announced the $3 billion, billion with A B, we 644 00:33:34,660 --> 00:33:38,380 don't do that too often, $3 billion investment in our in 645 00:33:38,380 --> 00:33:39,940 data and artificial intelligence. 646 00:33:40,640 --> 00:33:42,880 You know there's a good part of that is for generative A I, but 647 00:33:42,880 --> 00:33:44,680 it's across data and artificial intelligence. 648 00:33:44,680 --> 00:33:48,160 So we're we're doubling our workforce we have 40,000 people 649 00:33:48,160 --> 00:33:51,240 that work in data and a I today we do a lot of work in the area. 650 00:33:51,240 --> 00:33:52,760 We're going to double that over three years. 651 00:33:53,400 --> 00:33:57,440 We're going to we're developing a new tool called a I Navigator 652 00:33:57,440 --> 00:34:01,280 for enterprise to help companies apply a I more quickly including 653 00:34:01,280 --> 00:34:04,560 generative A I tool self uses uses generative A I to help 654 00:34:04,800 --> 00:34:07,280 companies understand the road map they need to to follow and 655 00:34:07,280 --> 00:34:09,679 how they you know, industry by industry, how they can drive 656 00:34:09,679 --> 00:34:12,760 value from. A I and and we we're creating a 657 00:34:12,760 --> 00:34:14,760 Center for advanced A I where we're looking not just a 658 00:34:14,760 --> 00:34:17,760 generative A I, but the next breakthroughs that will come as 659 00:34:18,080 --> 00:34:19,800 as well. So we're excited about it. 660 00:34:19,800 --> 00:34:22,920 We're putting, we're putting a lot of money and focus on it 661 00:34:22,920 --> 00:34:25,400 because we do believe this is transformational for business 662 00:34:25,400 --> 00:34:28,840 and this will this will this wave will build faster than 663 00:34:28,840 --> 00:34:31,840 cloud and faster than some of the other technology waves that 664 00:34:31,840 --> 00:34:36,440 we've that we've seen before. So yeah, so a big, big focus 665 00:34:36,520 --> 00:34:39,130 and. And we see companies doing the 666 00:34:39,130 --> 00:34:41,850 same. So we did a survey recently and 667 00:34:41,969 --> 00:34:45,409 97% of executives that we surveyed, this is just a couple 668 00:34:45,409 --> 00:34:49,810 weeks ago, 97% believe this is going to be strategic for their 669 00:34:49,810 --> 00:34:52,370 companies and and it's going to change their business or their 670 00:34:52,370 --> 00:34:57,050 industry. 97% that's basically everybody over 50% believe it's 671 00:34:57,050 --> 00:34:59,930 game changing, you know not just change some change but game 672 00:34:59,930 --> 00:35:01,370 changing for their industry or company. 673 00:35:01,810 --> 00:35:06,050 About 46% are going to invest a significant part of their budget 674 00:35:06,050 --> 00:35:09,410 and generative A I. In the next two years, this is a 675 00:35:09,410 --> 00:35:11,690 fast build and maybe some of this is you know, companies 676 00:35:11,690 --> 00:35:15,530 getting a little overexcited, but but we believe that that 677 00:35:15,610 --> 00:35:19,010 that pattern will hold and companies will move and invest 678 00:35:19,010 --> 00:35:21,650 in this technology, you know get more quickly than we've seen 679 00:35:21,890 --> 00:35:25,010 other ways of technology built. But what about the risk 680 00:35:25,010 --> 00:35:29,570 associated with investing in something where the end 681 00:35:29,810 --> 00:35:34,230 trajectory is so unclear? You need to look at the horizon. 682 00:35:34,230 --> 00:35:36,310 I think there's a lot of things that that are clear that you 683 00:35:36,310 --> 00:35:39,630 can, you know that are clear. I think the key thing is to look 684 00:35:39,630 --> 00:35:44,070 at this from 2 dimensions, business case, dimension and the 685 00:35:44,070 --> 00:35:47,230 response responsible A I to mention which helps you balance 686 00:35:47,230 --> 00:35:48,510 the risk. The business case helps you look 687 00:35:48,510 --> 00:35:51,270 at the value responsible. A I helps you look at you know 688 00:35:51,350 --> 00:35:54,070 applying with the human values and right risk profile. 689 00:35:54,270 --> 00:35:56,670 If you take two those two lenses, I think you can find the 690 00:35:56,670 --> 00:35:58,030 intersection of the right things. 691 00:35:58,270 --> 00:36:00,990 You can start on now with no regrets, and obviously you have 692 00:36:00,990 --> 00:36:02,910 to make sure that the use case you look at. 693 00:36:03,110 --> 00:36:05,510 As you know can be supported with the technology that's 694 00:36:05,510 --> 00:36:07,830 available today which is moving super fast. 695 00:36:07,870 --> 00:36:12,230 So I think I think Michael we can you can identify no regrets 696 00:36:12,230 --> 00:36:14,830 things to do. We believe in the near term this 697 00:36:14,830 --> 00:36:17,550 is going to be human in the loop types of solutions for the for 698 00:36:17,550 --> 00:36:19,910 the most part it's going to be solutions that bring in 699 00:36:19,910 --> 00:36:21,990 tremendous new capabilities for people. 700 00:36:22,230 --> 00:36:25,030 It's going to be, you know, new exciting capabilities for 701 00:36:25,030 --> 00:36:29,750 consumers to use, you know, more directly like in in one case, a 702 00:36:29,750 --> 00:36:32,270 retailer we're working with that's using generative A I to 703 00:36:32,270 --> 00:36:35,230 create all sorts. Of new product configuration 704 00:36:35,230 --> 00:36:39,110 capability for their customer, for their customers, it's going 705 00:36:39,110 --> 00:36:41,070 to create new capability for employees etcetera. 706 00:36:41,190 --> 00:36:43,630 This is all stuff that's doable today, I think, with no regrets, 707 00:36:43,630 --> 00:36:46,670 without, you know, worrying too much about the risk and you can 708 00:36:46,750 --> 00:36:49,790 you can apply the right principles to to do it in a 709 00:36:49,790 --> 00:36:52,920 responsible way. From an industry specific 710 00:36:52,920 --> 00:36:56,520 standpoint, it seems like each industry is dealing with a I at 711 00:36:56,600 --> 00:36:59,880 it at its own speed. The two that I want to bring up 712 00:36:59,880 --> 00:37:02,840 right now that I've had probably, I would say some of 713 00:37:02,840 --> 00:37:06,640 the most impact is 1 education and two, the legal sector. 714 00:37:06,720 --> 00:37:09,760 The funny thing about it is they dealt with this regulation in 715 00:37:09,760 --> 00:37:12,120 entirely different ways. In the education sector, 716 00:37:12,320 --> 00:37:14,760 everything is pretty much a chaotic mess. 717 00:37:14,760 --> 00:37:17,720 You know, you have schools banning things, turning things 718 00:37:17,720 --> 00:37:20,440 off, then re enabling and and there's a whole that we could 719 00:37:20,440 --> 00:37:23,160 have a whole show. On this but on the legal side, 720 00:37:23,320 --> 00:37:27,040 you've got which surprises me the most as a technologies 721 00:37:27,640 --> 00:37:29,800 lawyers really embracing this technology. 722 00:37:30,200 --> 00:37:32,640 There's obviously a little resentment but there's there's 723 00:37:32,680 --> 00:37:36,480 legal LL Ms. and there's a lot of adoption as to how you can 724 00:37:36,480 --> 00:37:39,480 integrate it and adopt it and and and make you know your law 725 00:37:39,480 --> 00:37:41,160 firm or your practice move faster. 726 00:37:41,360 --> 00:37:44,080 I would have never predicted that in 100 years but it's 727 00:37:44,080 --> 00:37:46,840 happening now. On the flip side, Elizabeth Shaw 728 00:37:46,840 --> 00:37:49,680 from Twitter has this really good question where a lot of 729 00:37:49,680 --> 00:37:53,000 organizations and individuals. Have begun using generative AI 730 00:37:53,000 --> 00:37:56,560 for work without any AI governance in place and she's 731 00:37:56,560 --> 00:37:58,960 wondering how you can apply governance once the horses are 732 00:37:58,960 --> 00:38:01,720 out of the barn and racing. The reason why I brought up the 733 00:38:01,840 --> 00:38:05,520 the points earlier is you know education, that whole sector is 734 00:38:05,520 --> 00:38:08,880 dealing with this, this, this whole dilemma right now. 735 00:38:09,160 --> 00:38:12,200 And I'm curious your take just because you're seeing it on the 736 00:38:12,200 --> 00:38:16,430 enterprise side where? If I input, you know, an e-mail 737 00:38:16,430 --> 00:38:19,710 or consonants of a document, there's a true risk there, 738 00:38:19,710 --> 00:38:21,950 whether it be IP or trade secrets. 739 00:38:22,110 --> 00:38:26,070 Whereas with school, you know if I put my quiz test is quiz 740 00:38:26,070 --> 00:38:29,550 questions and test questions in the program, you know it really 741 00:38:29,550 --> 00:38:33,870 only impacts me and and and will have a lasting impact on the the 742 00:38:33,870 --> 00:38:35,430 knowledge that I retain and gain. 743 00:38:35,630 --> 00:38:38,790 We're seeing broad adoption across industries unlike other 744 00:38:38,790 --> 00:38:41,470 any other technology I've seen which had very specific and 745 00:38:41,710 --> 00:38:43,510 everything had specific industry patterns. 746 00:38:43,750 --> 00:38:49,110 Client, server, ERP, mobility, Cloud SASS had very specific 747 00:38:49,110 --> 00:38:51,590 industry patterns. Generative A I is super broad in 748 00:38:51,590 --> 00:38:55,150 terms of the industry adoption we're seeing in the industry, 749 00:38:55,150 --> 00:38:57,750 potential use cases we're seeing the two you mentioned are are 750 00:38:57,750 --> 00:39:00,510 super interesting. CUE education I think will be 751 00:39:00,790 --> 00:39:03,990 literally transformed and and you know through generative A I, 752 00:39:03,990 --> 00:39:07,070 it enables truly personalized learning in ways that are 753 00:39:07,270 --> 00:39:09,430 significantly different than our current educational system. 754 00:39:09,430 --> 00:39:11,870 It'll take a while for that to work through, but yes, it's 755 00:39:11,870 --> 00:39:13,670 going to be, you know, pervasive and powerful. 756 00:39:13,750 --> 00:39:15,990 Legal I agree with you. The interesting thing about the 757 00:39:15,990 --> 00:39:19,790 legal profession is it can help paralegals work more effectively 758 00:39:19,790 --> 00:39:22,190 and do higher level work and it could allow experienced lawyers 759 00:39:22,390 --> 00:39:24,710 to lever to themselves more effectively in terms of the work 760 00:39:24,710 --> 00:39:26,670 they get done. So we're seeing it being adopted 761 00:39:26,670 --> 00:39:29,350 across you know the different types of work in the in the 762 00:39:29,350 --> 00:39:31,430 legal. Industry or legal profession 763 00:39:31,430 --> 00:39:34,990 from that perspective but I think to the horse out of the 764 00:39:34,990 --> 00:39:38,030 barn question. Yeah you can you can still apply 765 00:39:38,070 --> 00:39:41,150 responsible A I you can go back through and and do it. 766 00:39:41,350 --> 00:39:43,230 It's a matter of being systematic and rigorous. 767 00:39:43,230 --> 00:39:45,470 It's about having C-Suite and CEO support. 768 00:39:45,710 --> 00:39:47,910 We report on responsible A I to our board. 769 00:39:48,230 --> 00:39:52,230 It's part of our formal compliance responsibility that 770 00:39:52,230 --> 00:39:54,430 we do and we encourage organizations to do the same. 771 00:39:54,790 --> 00:39:58,670 And and if you already have a I out there and most organizations 772 00:39:58,670 --> 00:40:00,810 do. And most organizations don't 773 00:40:00,810 --> 00:40:02,330 have enough response by a I in place. 774 00:40:02,890 --> 00:40:06,210 We believe it's time to do that inventory the A I know where you 775 00:40:06,210 --> 00:40:08,050 using it. Understand the risk level, know 776 00:40:08,050 --> 00:40:10,970 the mediation techniques and tools and have them at at your 777 00:40:10,970 --> 00:40:13,650 disposal and know if you've mediated the risks. 778 00:40:13,930 --> 00:40:17,730 You have to go back right track to do that if you haven't done 779 00:40:17,730 --> 00:40:19,690 it so that you know what your baseline is as you start to 780 00:40:19,690 --> 00:40:22,770 apply more A I in generative A I going forward. 781 00:40:23,370 --> 00:40:29,090 Given the impact of AI, we know that it will be pro it is 782 00:40:29,090 --> 00:40:34,690 profound and will be profound. Where is this going? 783 00:40:34,690 --> 00:40:39,130 And more importantly, how should businesses position themselves 784 00:40:39,130 --> 00:40:44,250 to capitalize on this obvious sea change that's kind of 785 00:40:44,250 --> 00:40:47,980 erupting all around us? I think this the simple answer 786 00:40:47,980 --> 00:40:50,820 is you need to think big, start small and scale fast. 787 00:40:50,820 --> 00:40:54,340 But think big is think about where this what the real 788 00:40:54,340 --> 00:40:57,740 potential is and where this could you take your organization 789 00:40:57,740 --> 00:40:59,900 where the the big threats the big opportunities that's 790 00:40:59,900 --> 00:41:01,540 thinking big. Start small. 791 00:41:01,540 --> 00:41:04,420 Small as the experiment with the human in the loop and the no 792 00:41:04,420 --> 00:41:06,580 regrets use cases. Get some experience, understand 793 00:41:06,580 --> 00:41:08,300 the models. Select the right partners, your 794 00:41:08,300 --> 00:41:12,180 models and such and and and do something and get ready to scale 795 00:41:12,180 --> 00:41:13,300 fast. This is the centers of 796 00:41:13,300 --> 00:41:16,200 excellence. The the operational maturity 797 00:41:16,200 --> 00:41:19,160 that's one of the good questions came in on and another 798 00:41:19,160 --> 00:41:21,400 capabilities that the talent that you built around it to 799 00:41:21,400 --> 00:41:23,840 scale fast. So that's say I think big start 800 00:41:23,840 --> 00:41:25,960 small scale fast. It's a good advice on give. 801 00:41:26,570 --> 00:41:30,090 Is scifi has shown us, you know, what the future has looked like. 802 00:41:30,090 --> 00:41:33,410 I mean we see some of the gadgets and gizmos that are real 803 00:41:33,410 --> 00:41:38,290 life objects from Star Trek. We see some of the unforeseen 804 00:41:38,330 --> 00:41:41,610 and uncomfortable futures from Black Mirror start to arise. 805 00:41:41,890 --> 00:41:44,450 One of the things that I'm wondering in your take is, I 806 00:41:44,450 --> 00:41:47,490 mean you wrote the book Human and Plus Machine and then you've 807 00:41:47,490 --> 00:41:50,610 got another one. Since then, I'm guessing you've 808 00:41:50,610 --> 00:41:54,410 been thinking about this whole concept of transhumanism and 809 00:41:54,690 --> 00:41:57,890 merging, kind of the the. The brain computer interfaces 810 00:41:57,890 --> 00:42:02,050 that Elon talks about with some of these AI models, like how how 811 00:42:02,050 --> 00:42:04,050 near do you think that is? Or do you think that that is 812 00:42:04,050 --> 00:42:07,370 still fodder for science fiction novelists? 813 00:42:07,810 --> 00:42:10,370 First of all, I'm a massive fan of science fiction, and I 814 00:42:10,370 --> 00:42:12,730 believe most science fiction eventually becomes real. 815 00:42:12,730 --> 00:42:17,210 It's a matter of the timeline and if you want to, if you want 816 00:42:17,290 --> 00:42:19,430 to. Read about where technology is 817 00:42:19,430 --> 00:42:20,750 going. You pick up, you know, somebody 818 00:42:20,750 --> 00:42:24,030 like a Neil Stephenson and read his books where he, he's 819 00:42:24,030 --> 00:42:26,630 anticipated coin, the term metaverse, among other things, 820 00:42:26,950 --> 00:42:30,190 and his befall previewed where we are with technology right 821 00:42:30,190 --> 00:42:32,470 now, you know, a number of years ago really well. 822 00:42:33,190 --> 00:42:35,830 So science fiction can be incredibly illuminating into 823 00:42:35,830 --> 00:42:37,830 where we're going in terms of transhumanism. 824 00:42:38,190 --> 00:42:41,150 You know, I'm not a real expert per se in that field, but I 825 00:42:41,470 --> 00:42:45,270 talked to a lot of friends and colleagues who are, and I I 826 00:42:45,270 --> 00:42:47,550 believe it's quite far away. I mean, think about. 827 00:42:47,910 --> 00:42:50,110 Blown away. We are by large language models 828 00:42:50,110 --> 00:42:51,870 today. And ChatGPT and everything. 829 00:42:52,190 --> 00:42:54,750 There is no intelligence inherent in these models. 830 00:42:55,030 --> 00:42:56,950 These are, these are, these are statistical models. 831 00:42:56,950 --> 00:42:59,110 So people ask me how intelligent these models are. 832 00:42:59,150 --> 00:43:02,650 The models have no intelligence. The models are a bunch of data 833 00:43:02,650 --> 00:43:05,930 with the technology that can you know, that can statistically 834 00:43:05,930 --> 00:43:08,690 create results from them. There is no inherent knowledge. 835 00:43:09,010 --> 00:43:11,970 Now, some of the breakthroughs we're looking for in AI, the 836 00:43:11,970 --> 00:43:15,170 next generations of things like common sense AI, the way 837 00:43:15,170 --> 00:43:18,330 knowledge graphs come in and can be combined with generative AI 838 00:43:18,530 --> 00:43:21,910 that starts to create, you know. Systems that have more 839 00:43:21,910 --> 00:43:25,630 intelligence, you know, inherent in the models along with the 840 00:43:25,630 --> 00:43:28,030 generative capability. And I think that's where you see 841 00:43:28,030 --> 00:43:31,830 some interesting advances. But truly getting to the human, 842 00:43:32,550 --> 00:43:35,750 you know, human and and surpassing human level, you 843 00:43:35,750 --> 00:43:38,350 know, I think, I think we're quite far away from it. 844 00:43:39,470 --> 00:43:42,070 And I were multiple, you know, multiple of multiple 845 00:43:42,070 --> 00:43:45,030 breakthroughs, you know, away from, I believe from from. 846 00:43:45,030 --> 00:43:47,460 Seeing that, I think. Look, that discussion I think, 847 00:43:47,460 --> 00:43:50,180 distracts us a little bit from what we need to do today, which 848 00:43:50,180 --> 00:43:53,460 is some of the great questions that listeners have asked about 849 00:43:53,700 --> 00:43:56,100 human values and ethics and of what? 850 00:43:56,100 --> 00:43:57,980 You know, let's let's prevent people from using today's 851 00:43:57,980 --> 00:44:01,580 technology in bad ways and and avoid getting a little bit too 852 00:44:01,580 --> 00:44:03,820 distracted by the things that are that are pretty far down the 853 00:44:03,820 --> 00:44:07,060 road. This is from Mike Prest. 854 00:44:07,060 --> 00:44:09,460 He's a chief information officer on LinkedIn. 855 00:44:09,460 --> 00:44:14,180 He says, as a business leader managing the risks of AI, how do 856 00:44:14,180 --> 00:44:16,340 you? What advice can you offer on 857 00:44:16,340 --> 00:44:21,180 sharing information to become good stewards of the technology 858 00:44:21,500 --> 00:44:26,860 and dispel some of the dystopian conversations about generative 859 00:44:27,100 --> 00:44:28,580 AI? And very quickly, please. 860 00:44:28,940 --> 00:44:31,660 I think we should share more. I'm happy to on that front. 861 00:44:31,660 --> 00:44:34,220 I'm happy to you know connect with anybody and share some 862 00:44:34,220 --> 00:44:35,900 ideas. There's some various forms out 863 00:44:35,900 --> 00:44:38,860 there where there's a lot of the sharing happening both in 864 00:44:38,860 --> 00:44:40,740 business communities and different technology for them. 865 00:44:40,740 --> 00:44:42,580 So I think that's how they'll all get better. 866 00:44:42,580 --> 00:44:45,740 It's it's it's at the early stages that I'm I have a lot of 867 00:44:45,740 --> 00:44:48,660 forms that I'm running with some of my peers and colleagues and 868 00:44:48,660 --> 00:44:51,740 and other companies to to to share a lot because we're all 869 00:44:51,740 --> 00:44:53,980 learning together in this fast moving technology. 870 00:44:54,660 --> 00:44:58,700 And we have another question from Twitter another really good 871 00:44:58,700 --> 00:45:00,700 one and again really quickly please. 872 00:45:00,980 --> 00:45:04,580 And this is from James McGovern who says with Microsoft and 873 00:45:04,580 --> 00:45:07,980 Oracle holding layoffs the talent for enterprise 874 00:45:07,980 --> 00:45:12,660 architecture and sales professionals must be huge. 875 00:45:12,900 --> 00:45:15,460 Who's hiring? Enterprise architecture. 876 00:45:15,540 --> 00:45:18,340 You know, as much as the generative AI skills enterprise 877 00:45:18,340 --> 00:45:21,670 architecture is. Immensely important Generative A 878 00:45:21,670 --> 00:45:24,870 I is creating and and along with the Metaverse capabilities which 879 00:45:24,870 --> 00:45:28,550 we didn't talk about in this this call it, it creates really 880 00:45:28,550 --> 00:45:30,830 a rethink of your enterprise architecture what you need to 881 00:45:30,830 --> 00:45:31,990 do. So those skills I think are 882 00:45:31,990 --> 00:45:34,110 tremendous demand as we look at this going forward. 883 00:45:34,550 --> 00:45:37,310 So I think a lot of companies are looking at hiring the right 884 00:45:37,310 --> 00:45:40,030 talent to to build this out. I think enterprise architects in 885 00:45:40,030 --> 00:45:42,870 in particular has been a shortage in the industry for a 886 00:45:42,870 --> 00:45:45,590 while in our you know even even more demand as we go forward 887 00:45:45,590 --> 00:45:48,430 with the every every new technology like generative A I 888 00:45:49,080 --> 00:45:54,560 Paul let's shift gears here you're an avid sailor I've known 889 00:45:54,560 --> 00:46:00,840 you for many years and and I see you sailing tell us why. 890 00:46:01,360 --> 00:46:04,680 What do you tell us about your sailing and and why do you like 891 00:46:04,680 --> 00:46:07,680 to sail so much? I've sailed my whole life, so 892 00:46:07,720 --> 00:46:13,560 it's something that's been a lifelong passion and I find I 893 00:46:13,920 --> 00:46:17,160 think I love the experience of it when you're out on the on the 894 00:46:17,160 --> 00:46:19,660 water. And you're seeing the sunset. 895 00:46:19,660 --> 00:46:22,940 You have a nice breeze behind you and you're and you're 896 00:46:22,940 --> 00:46:27,060 powered only by the wind and it's sailing along and can hear 897 00:46:27,060 --> 00:46:30,140 the little bubbling under the keel of your boat as you're as 898 00:46:30,140 --> 00:46:31,700 you're moving through the water at a nice pace. 899 00:46:31,700 --> 00:46:34,060 There's not a better feeling in the world than that. 900 00:46:34,060 --> 00:46:37,060 There's a, there's a challenge aspect of it, which is. 901 00:46:38,170 --> 00:46:41,170 Which is optimizing. How do you go a little faster? 902 00:46:41,170 --> 00:46:42,930 How do you get the sales tuned a little bit better? 903 00:46:42,930 --> 00:46:45,810 And I love the intellectual challenge of that. 904 00:46:46,050 --> 00:46:48,010 There's a learning aspect I learned something about. 905 00:46:48,250 --> 00:46:51,290 I've been selling my whole life. I learned something new, either 906 00:46:51,530 --> 00:46:54,330 by making a mistake or just encountering something every 907 00:46:54,330 --> 00:46:57,130 single time I'm on the boat. And it's as a continual, 908 00:46:57,290 --> 00:47:00,770 continual learning experience. And finally, I just say it's 909 00:47:00,890 --> 00:47:03,910 it's my happy place. You know the one, it's the one 910 00:47:03,910 --> 00:47:06,750 place where I really don't think about anything else because from 911 00:47:06,750 --> 00:47:09,550 a safety perspective and focusing on what what I'm doing 912 00:47:09,550 --> 00:47:12,350 on the boat and everything else. When I'm when I'm on my boat, 913 00:47:12,350 --> 00:47:14,950 that's that's where I am and that's where my whole focus in 914 00:47:14,950 --> 00:47:18,070 my mind is that is on my, you know, on my boat and the the 915 00:47:18,070 --> 00:47:19,990 guests and passengers that I that I have on it. 916 00:47:20,430 --> 00:47:23,870 As an author, I'm sure your some of your pastime includes 917 00:47:23,870 --> 00:47:26,270 reading. What books are you reading these 918 00:47:26,270 --> 00:47:27,950 days, and and what's keeping you sane? 919 00:47:28,500 --> 00:47:31,620 One of my favorite authors and heroes is Neil Stephenson, who's 920 00:47:31,620 --> 00:47:36,620 wrote so many great science science fiction books, so I put 921 00:47:36,620 --> 00:47:39,180 him out there. A great book that I read 922 00:47:39,380 --> 00:47:44,100 recently is Cloud Atlas, which is a fantastic story that gets 923 00:47:44,100 --> 00:47:46,100 into some of the top topics that we talked about. 924 00:47:46,100 --> 00:47:49,420 It's a prize winning novel that covers everything from the fall 925 00:47:49,540 --> 00:47:53,100 of the Ottoman Empire to to space travel in the future. 926 00:47:54,070 --> 00:47:56,230 In through a series of parallel stories. 927 00:47:56,230 --> 00:47:59,750 So it's a very interesting read. There's a book called Reality 928 00:47:59,750 --> 00:48:03,910 Plus which is I'd recommend to anybody anyone that's interested 929 00:48:04,350 --> 00:48:05,990 in it. Well first of all, transhumanism 930 00:48:05,990 --> 00:48:08,710 topic. You mentioned the metaverse or 931 00:48:08,830 --> 00:48:11,830 or related topics. Reality Plus is by a philosopher 932 00:48:11,830 --> 00:48:15,390 from NYU is exploring the question of are we living in a 933 00:48:15,390 --> 00:48:18,470 real world or simulation and how do you know the difference 934 00:48:18,470 --> 00:48:20,230 between the two. It's a fascinating book and 935 00:48:20,510 --> 00:48:25,740 super well written so. I I read a lot and it's it's 936 00:48:25,940 --> 00:48:28,660 those give you a sense of the the realm of from fiction to 937 00:48:28,660 --> 00:48:31,700 science fiction to you know kind of philosophy as well as 938 00:48:31,700 --> 00:48:35,440 technology. You're the senior person for 939 00:48:35,440 --> 00:48:41,640 technology at Accenture which employs about 740,000 people. 940 00:48:41,840 --> 00:48:44,280 I mean, just that number in and of itself is almost 941 00:48:44,480 --> 00:48:46,960 incomprehensible. How do you spread yourself over 942 00:48:46,960 --> 00:48:52,160 740,000 people and manage the pressure and the expectations? 943 00:48:52,480 --> 00:48:56,080 It's an amazing privilege, you know that to have a role like 944 00:48:56,080 --> 00:48:58,000 this because we our mission is to. 945 00:48:59,080 --> 00:49:01,920 Deliver on the promise of technology and human ingenuity. 946 00:49:02,120 --> 00:49:03,720 And the human. Ingenuity that we have the 947 00:49:03,720 --> 00:49:10,080 740,000 people is just amazing and what I what I what I like 948 00:49:10,080 --> 00:49:13,840 most about my job is able to learn from 740,000 people. 949 00:49:13,840 --> 00:49:16,240 I don't talk to each of them individually but the work that 950 00:49:16,240 --> 00:49:18,720 we do for clients, the innovative ideas they come up 951 00:49:18,720 --> 00:49:22,160 with is just super inspiring. You know the ways that they we 952 00:49:22,160 --> 00:49:25,240 that the projects we do in terms of improving communities and 953 00:49:25,240 --> 00:49:27,640 society through some of the work we do so. 954 00:49:28,620 --> 00:49:31,620 So it's it's really a privilege to to to do it and I'm just 955 00:49:31,620 --> 00:49:36,580 honored to have the role and to and to represent represent the 956 00:49:36,620 --> 00:49:39,620 the amazing group of people that we have the amazing leadership 957 00:49:40,460 --> 00:49:44,220 that we have And you know it it sounds it is a big company it's 958 00:49:44,220 --> 00:49:47,620 a lot of people but it's it's a lot of small communities that 959 00:49:47,620 --> 00:49:49,940 come together with a common culture is the way to think 960 00:49:49,940 --> 00:49:53,100 about is the way to think about it and we we have the the system 961 00:49:53,100 --> 00:49:55,660 we know how to hire you know people in volume if we need to 962 00:49:55,660 --> 00:49:59,700 we we know we know how to. We build community and build 963 00:49:59,700 --> 00:50:02,580 culture in our organization in a lot of different ways. 964 00:50:02,580 --> 00:50:06,620 So some things, you know as you scale up and get bigger, some 965 00:50:06,620 --> 00:50:09,900 things aren't that much harder to do at bigger scale and ended 966 00:50:09,900 --> 00:50:12,020 up scaling, you know very well as you grow and that's what I've 967 00:50:12,180 --> 00:50:13,220 know. What I've found is we've grown 968 00:50:13,220 --> 00:50:16,260 the organization. So it's a it's a it's a lot of 969 00:50:16,260 --> 00:50:17,340 fun. And it again, it's just a 970 00:50:17,540 --> 00:50:20,060 privilege to, you know, be be in an organization like this and 971 00:50:20,060 --> 00:50:22,460 have the role that I have. What's the hardest part? 972 00:50:22,820 --> 00:50:25,420 I don't know all the 740,000 names, but I'm working my way 973 00:50:25,420 --> 00:50:28,480 through as best I can. Hey Paul, question for you 974 00:50:28,480 --> 00:50:33,360 regarding just being a techie. What's your favorite device? 975 00:50:33,800 --> 00:50:36,440 Probably apps that I use. So what? 976 00:50:36,440 --> 00:50:39,000 What are the device I'm really getting a kick out of is my aura 977 00:50:39,000 --> 00:50:41,500 ring. Not really marketing for a 978 00:50:41,500 --> 00:50:44,940 specific product, but it's simple device, It's the ring, 979 00:50:44,940 --> 00:50:47,340 it's connected the app on the phone, and I'm finding it's 980 00:50:47,340 --> 00:50:51,740 really helping me understand some patterns on how I can be a 981 00:50:51,740 --> 00:50:54,620 little healthier and happier and get better sleep and such. 982 00:51:01,740 --> 00:51:05,260 And compared to my sleep activity, compared to my sleep 983 00:51:05,260 --> 00:50:56,340 cycle. You can track and I can track 984 00:50:56,340 --> 00:50:59,980 and correlate my heart rate, my oxygenation, my breathing 985 00:50:59,980 --> 00:51:07,100 patterns, all sorts of. Compared to my activity cycle 986 00:51:07,420 --> 00:51:10,260 and it's we're data-driven you know and if you get better data 987 00:51:10,260 --> 00:51:11,980 you can improve patterns and such. 988 00:51:11,980 --> 00:51:13,940 That's one of my one of the things I'm playing around with 989 00:51:13,940 --> 00:51:16,780 the right now that I'm I'm getting getting a lot of value 990 00:51:16,780 --> 00:51:18,500 out of. So one of the things that's 991 00:51:18,500 --> 00:51:22,540 interesting about the Aura ring is it represents like the whole 992 00:51:22,540 --> 00:51:25,620 quantified self movement. So you now have your own 993 00:51:25,620 --> 00:51:28,940 personal database of data that you can do whatever you want 994 00:51:28,940 --> 00:51:30,420 with. Are you going to build anything 995 00:51:30,420 --> 00:51:33,700 using your health data or is this just a personal experience? 996 00:51:33,980 --> 00:51:37,060 I don't know but I I'm on that exact journey you mentioned, I'm 997 00:51:37,060 --> 00:51:39,420 starting with now the by the personal bio understanding that 998 00:51:39,420 --> 00:51:43,100 your bio more using the you know the self you know diagnostics 999 00:51:43,100 --> 00:51:45,860 you can which has this another big impact on on. 1000 00:51:47,080 --> 00:51:49,160 Kind of health and Wellness. So yeah, I've been trying to get 1001 00:51:49,160 --> 00:51:51,480 more and more kind of data-driven and understanding, 1002 00:51:51,720 --> 00:51:54,200 you know, kind of what what, what makes me work and what 1003 00:51:54,200 --> 00:51:57,120 makes me healthy or not. So yeah, that's that is 1004 00:51:57,120 --> 00:51:58,240 something I'm going to continue doing. 1005 00:51:58,680 --> 00:52:00,960 It's funny because that's the big data that comes off of your 1006 00:52:00,960 --> 00:52:04,160 body and then you take that what works for you and implement that 1007 00:52:04,160 --> 00:52:08,440 at the enterprise at scale. I see what you're doing exactly 1008 00:52:09,280 --> 00:52:13,360 OK and with that we are out of time. 1009 00:52:13,560 --> 00:52:16,760 A huge thank you to Paul Doherty. 1010 00:52:16,760 --> 00:52:20,360 He's the Chief executive for Accenture Technology. 1011 00:52:20,360 --> 00:52:23,880 Paul, thank you for coming back again to CXO Talk. 1012 00:52:23,880 --> 00:52:25,560 We really, really do appreciate it. 1013 00:52:25,560 --> 00:52:27,120 It. Was a pleasure, Michael, and 1014 00:52:27,120 --> 00:52:28,880 it's great to do this with that with Q as well. 1015 00:52:28,880 --> 00:52:31,400 So thanks thanks to you both. And to the audience, those are 1016 00:52:31,400 --> 00:52:33,640 amazing questions. I wish I could be there and ask 1017 00:52:33,640 --> 00:52:35,960 the audience a lot of questions as well, but it's a pretty great 1018 00:52:35,960 --> 00:52:38,760 experience. Thank you and Q Harrison, it's 1019 00:52:38,760 --> 00:52:42,200 great to see you and and thank you for being such a great 1020 00:52:42,360 --> 00:52:43,760 cohost. That was a lot of fun, wasn't 1021 00:52:43,760 --> 00:52:45,320 it, Q? Indeed, man. 1022 00:52:45,320 --> 00:52:47,160 Thank you for having me everybody. 1023 00:52:47,280 --> 00:52:49,800 Thank you for watching. And as Paul said, you guys are 1024 00:52:49,800 --> 00:52:53,880 an amazing audience. Before you go, be sure to 1025 00:52:53,880 --> 00:52:58,560 subscribe to our newsletter, subscribe to our YouTube 1026 00:52:58,560 --> 00:53:02,520 channel, check out cxotalk.com and we will see you again next 1027 00:53:02,520 --> 00:53:04,200 time. We have amazing, really great 1028 00:53:04,200 --> 00:53:05,760 shows coming up. Have a great day everybody. 1029 00:53:05,840 --> 00:53:06,320 Bye, bye.