1 00:00:02,440 --> 00:00:06,760 Speaker 1: Bloomberg Audio Studios, podcasts, radio News. 2 00:00:07,000 --> 00:00:10,319 Speaker 2: I'm joined by Alex Burkhart, UK Cabinet Office Minister leading 3 00:00:10,360 --> 00:00:13,039 Speaker 2: on AI in the public sector. I want to start 4 00:00:13,039 --> 00:00:15,680 Speaker 2: out with your role in artificial intelligence. This role wouldn't 5 00:00:15,720 --> 00:00:18,520 Speaker 2: exist even as two or three years ago, and need 6 00:00:18,560 --> 00:00:21,079 Speaker 2: to live in the United Kingdom. We've seen this government, 7 00:00:21,160 --> 00:00:23,759 Speaker 2: led by the Prime Minister Richie Sunek, really want to 8 00:00:23,840 --> 00:00:27,479 Speaker 2: invest and hone AI even within the government. I know 9 00:00:27,560 --> 00:00:30,720 Speaker 2: you're in the beta mode of this red box. Tell 10 00:00:30,760 --> 00:00:32,960 Speaker 2: us a little bit about what you're doing, what you're 11 00:00:32,960 --> 00:00:35,960 Speaker 2: focused on, and how the government could potentially harness this. 12 00:00:36,320 --> 00:00:38,720 Speaker 3: Thanks Amory, thank you for having me on and it's 13 00:00:38,760 --> 00:00:40,280 Speaker 3: great to be in New York to talk about the 14 00:00:40,320 --> 00:00:42,120 Speaker 3: work we're doing in the UK government at the moment. 15 00:00:42,479 --> 00:00:45,480 Speaker 3: As you said, the Prime Minister is really passionate about AI. 16 00:00:45,600 --> 00:00:48,440 Speaker 3: We held the first AI safety summit that the world 17 00:00:48,600 --> 00:00:52,199 Speaker 3: seen last year, very very productive. But the work that 18 00:00:52,240 --> 00:00:54,760 Speaker 3: I'm doing is on how we can deploy AI for 19 00:00:54,800 --> 00:00:57,920 Speaker 3: the public good and that means how we can you know, 20 00:00:58,040 --> 00:01:01,560 Speaker 3: it's about eighteen months ago we sat down and said 21 00:01:01,720 --> 00:01:04,280 Speaker 3: this new technology is emerging potentially one of the great 22 00:01:04,360 --> 00:01:06,920 Speaker 3: leaps forward. How are we going to harness it to 23 00:01:06,959 --> 00:01:10,120 Speaker 3: improve the quality of public services and drive down costs 24 00:01:10,160 --> 00:01:12,760 Speaker 3: for the taxpayer. And there's a number of ways in 25 00:01:12,800 --> 00:01:15,720 Speaker 3: which that's emerging. And the first is that you can 26 00:01:15,760 --> 00:01:19,880 Speaker 3: now automate tasks like never before. And I have you 27 00:01:19,920 --> 00:01:22,600 Speaker 3: mentioned Red Box, which is named after the you know, 28 00:01:22,600 --> 00:01:25,279 Speaker 3: the red ministerial box that British ministers carry around. 29 00:01:25,480 --> 00:01:28,000 Speaker 1: I still hope they'll do that for show and the Yeah. 30 00:01:28,080 --> 00:01:32,200 Speaker 3: Absolutely. But what we now have, what we've been building 31 00:01:32,280 --> 00:01:36,640 Speaker 3: is a tool that can summarize huge documents very very quickly, 32 00:01:36,720 --> 00:01:39,240 Speaker 3: that can relate them to what's been going on in Parliament, 33 00:01:39,520 --> 00:01:42,400 Speaker 3: what's been going on in the press, and can pull 34 00:01:42,440 --> 00:01:45,000 Speaker 3: out kind of key points that ministers need to focus on. 35 00:01:45,360 --> 00:01:49,360 Speaker 3: And that's saving my Private office hundreds of hours. But 36 00:01:49,440 --> 00:01:52,280 Speaker 3: it's also starting to build a tool that is going 37 00:01:52,320 --> 00:01:57,160 Speaker 3: to create institutional memory. Recreate institutional memory in our government departments. 38 00:01:57,280 --> 00:01:59,760 Speaker 3: So you know, when we're starting out on a project, 39 00:01:59,760 --> 00:02:01,840 Speaker 3: we'll be able to ask, how we tried something like 40 00:02:01,880 --> 00:02:04,320 Speaker 3: this before, why didn't we carry on with it? You 41 00:02:04,360 --> 00:02:07,360 Speaker 3: know what was good about it? It's a huge resource 42 00:02:07,360 --> 00:02:13,200 Speaker 3: of documentation that's currently lying unused in government is going 43 00:02:13,240 --> 00:02:14,560 Speaker 3: to come back into play. 44 00:02:14,720 --> 00:02:17,520 Speaker 2: So two questions immediately come to mind, one talent and 45 00:02:17,560 --> 00:02:18,360 Speaker 2: the other security. 46 00:02:18,480 --> 00:02:19,760 Speaker 1: Let's start with talent first. 47 00:02:20,280 --> 00:02:23,040 Speaker 2: If you work in AI, you can get paid handsomely 48 00:02:23,200 --> 00:02:26,160 Speaker 2: going to say open AI. How do you attract that 49 00:02:26,280 --> 00:02:28,840 Speaker 2: talent to say, come work as a civil servant. 50 00:02:28,919 --> 00:02:31,240 Speaker 3: It's a great question. We were really worried about this 51 00:02:31,280 --> 00:02:34,799 Speaker 3: at the start because we can't pay that money. And 52 00:02:35,600 --> 00:02:38,920 Speaker 3: you know, I've got a really great head of the program, 53 00:02:39,000 --> 00:02:40,800 Speaker 3: Laura Gilbert, who's out in New York with me at 54 00:02:40,800 --> 00:02:45,000 Speaker 3: the moment, and she said, as long as we can 55 00:02:45,040 --> 00:02:48,519 Speaker 3: pay something, but more than we pay our normal civil servants. 56 00:02:48,800 --> 00:02:52,960 Speaker 3: What we have is data. We have really good data, 57 00:02:53,280 --> 00:02:56,360 Speaker 3: lots of centralized data. And if you are, you know, 58 00:02:56,440 --> 00:02:58,920 Speaker 3: an AI addict, if you're one of these super smart, 59 00:02:59,480 --> 00:03:02,200 Speaker 3: crazy CLIs people who wants to get involved, the thing 60 00:03:02,240 --> 00:03:05,280 Speaker 3: you want to play with is data. And so we've 61 00:03:05,600 --> 00:03:08,360 Speaker 3: we've been looking at huge quantities of data and things 62 00:03:08,400 --> 00:03:12,720 Speaker 3: like public prescriptions health service, looking how we can improve 63 00:03:13,240 --> 00:03:16,639 Speaker 3: the quality of prescriptions, drive down costs. We've been looking 64 00:03:17,080 --> 00:03:19,680 Speaker 3: at how we can eliminate fraud and error in the 65 00:03:19,680 --> 00:03:23,040 Speaker 3: benefit system. We have two hundred and eighty billion pound 66 00:03:23,720 --> 00:03:27,440 Speaker 3: benefit and pension system, and we currently employ thousands and 67 00:03:27,480 --> 00:03:29,840 Speaker 3: thousands of people to check for fraud. We think we're 68 00:03:29,880 --> 00:03:32,559 Speaker 3: going to be able to vastly reduce that and improve. 69 00:03:32,360 --> 00:03:34,200 Speaker 1: So attracting the talent hasn't been an issue. 70 00:03:34,320 --> 00:03:38,200 Speaker 3: No, No, we've we've been really pleasantly surprised that loads 71 00:03:38,240 --> 00:03:41,960 Speaker 3: of people have applied and they passed our very very 72 00:03:42,000 --> 00:03:43,040 Speaker 3: stringent tests to get in. 73 00:03:43,360 --> 00:03:44,920 Speaker 1: The one concern as well as security. 74 00:03:45,200 --> 00:03:48,760 Speaker 2: You put all of this now, documents that live in 75 00:03:48,800 --> 00:03:50,960 Speaker 2: a cloud, you're working more with AI. 76 00:03:51,480 --> 00:03:54,520 Speaker 1: How do you safeguard that to say adversaries? 77 00:03:54,760 --> 00:03:58,840 Speaker 3: Yeah, it's a central question and we have a very 78 00:03:58,840 --> 00:04:04,840 Speaker 3: effective security service in the UK and we're constantly learning 79 00:04:04,880 --> 00:04:08,560 Speaker 3: from partners. I was outing talent in Estonia right on 80 00:04:08,840 --> 00:04:12,040 Speaker 3: the fringes of NATO a few weeks ago, learning from 81 00:04:12,080 --> 00:04:15,520 Speaker 3: them because they were party to some enormous cyber attacks 82 00:04:15,520 --> 00:04:19,440 Speaker 3: a few years ago and have become very very resilient. 83 00:04:19,520 --> 00:04:22,360 Speaker 3: So it's uppermost in our mind when we're vetting people, 84 00:04:22,920 --> 00:04:25,320 Speaker 3: but also when we're building systems to make sure that 85 00:04:25,360 --> 00:04:30,120 Speaker 3: they're not you, that they can't be attacked by enemy operator. 86 00:04:30,160 --> 00:04:32,560 Speaker 2: If you use AI in the defense sector in the 87 00:04:32,640 --> 00:04:37,280 Speaker 2: UK government, how important is it that all that technology comes. 88 00:04:37,080 --> 00:04:41,039 Speaker 1: In house, not out of house. Well because of these concerns. 89 00:04:41,080 --> 00:04:44,960 Speaker 3: No, no, I think it's not just defense. Right when 90 00:04:45,000 --> 00:04:47,920 Speaker 3: we were starting out, we could easily have gone and 91 00:04:48,040 --> 00:04:50,839 Speaker 3: picked a number of big companies, big UK companies have 92 00:04:51,200 --> 00:04:55,760 Speaker 3: fantastic a sector in London. But we wanted to understand 93 00:04:55,760 --> 00:04:59,000 Speaker 3: this technology ourselves. So we've hired this cracked team of 94 00:04:59,120 --> 00:05:02,919 Speaker 3: thirty people going up to seventy who are going to 95 00:05:02,920 --> 00:05:05,400 Speaker 3: be able to build bespoke systems for us Now. In 96 00:05:05,440 --> 00:05:08,920 Speaker 3: the future, I think there'll be a case of some 97 00:05:08,960 --> 00:05:12,240 Speaker 3: stuff we will build ourselves and use many times, some 98 00:05:12,279 --> 00:05:17,200 Speaker 3: stuff we will build bespoke for particular departments, perhaps like defense, 99 00:05:17,760 --> 00:05:22,839 Speaker 3: and some things we will buy in. But having people 100 00:05:22,839 --> 00:05:26,040 Speaker 3: who actually know how it works and know what government's 101 00:05:26,080 --> 00:05:28,960 Speaker 3: needs are I think is central to getting this revolution right. 102 00:05:29,320 --> 00:05:30,919 Speaker 2: I know you're in New York, so you're missing all 103 00:05:31,000 --> 00:05:34,279 Speaker 2: the morning press in the UK after a critical election 104 00:05:34,680 --> 00:05:38,359 Speaker 2: last night and you saw the Bellwether parliamentary seat that 105 00:05:38,400 --> 00:05:42,159 Speaker 2: the Conservatives lost, do you honestly think the Conservative Party 106 00:05:42,160 --> 00:05:44,280 Speaker 2: if the elections are going to be held this year? 107 00:05:44,320 --> 00:05:45,960 Speaker 2: At some point I know the Prime Minister said at 108 00:05:45,960 --> 00:05:47,640 Speaker 2: the second half of the year, do you think the 109 00:05:47,680 --> 00:05:49,520 Speaker 2: current conservaives will be able to hold onto power. 110 00:05:49,880 --> 00:05:51,760 Speaker 3: I do think they will be able to and I'll 111 00:05:51,800 --> 00:05:54,880 Speaker 3: tell you right, look, we've had a tough night. The 112 00:05:54,920 --> 00:05:57,240 Speaker 3: truth is that in twenty twenty one, when this cycle 113 00:05:57,240 --> 00:06:00,760 Speaker 3: of elections was last on, we had a vaccine bounce 114 00:06:00,920 --> 00:06:02,880 Speaker 3: that was a very good year for us. We always 115 00:06:02,880 --> 00:06:06,719 Speaker 3: suspected we might come down. But there is a silver 116 00:06:06,760 --> 00:06:09,120 Speaker 3: lining to this cloud, and it's the fact that Ben 117 00:06:09,200 --> 00:06:12,720 Speaker 3: Houchin are mayor in Teesside, in the northeast, traditionally quite 118 00:06:12,760 --> 00:06:15,240 Speaker 3: a deprived part of the country, which under the Conservatives 119 00:06:15,279 --> 00:06:18,720 Speaker 3: is now flourishing. He's won, he's held on, He's done 120 00:06:18,760 --> 00:06:21,039 Speaker 3: a very good job there, and I think when it 121 00:06:21,080 --> 00:06:23,440 Speaker 3: comes to the election and people actually have to make 122 00:06:23,440 --> 00:06:25,159 Speaker 3: a choice about who they want to see in power 123 00:06:25,240 --> 00:06:27,680 Speaker 3: for the next five years. Is it going to be 124 00:06:27,760 --> 00:06:31,080 Speaker 3: the Prime Minister and the Conservatives that has got us 125 00:06:31,080 --> 00:06:35,200 Speaker 3: through COVID, that has seen off inflation and the energy 126 00:06:35,240 --> 00:06:38,120 Speaker 3: crisis and has a plan for the future, or is 127 00:06:38,160 --> 00:06:40,000 Speaker 3: it going to be a Labor party that really has 128 00:06:40,040 --> 00:06:42,359 Speaker 3: no plan and no agenda. I think they're going to 129 00:06:42,360 --> 00:06:42,840 Speaker 3: come with us. 130 00:06:43,360 --> 00:06:46,560 Speaker 2: Do now regret though potentially removing Boris Johnson. Just from 131 00:06:46,600 --> 00:06:49,320 Speaker 2: an electability point of view, No. 132 00:06:49,440 --> 00:06:54,480 Speaker 3: I don't because Boris certainly had his strengths, but things 133 00:06:54,480 --> 00:06:59,320 Speaker 3: weren't working out and he made some mistakes and we 134 00:06:59,320 --> 00:07:01,440 Speaker 3: couldn't have that those mistakes haven't happened. 135 00:07:01,640 --> 00:07:03,839 Speaker 2: Besides Ai, which I know the Prime Minister is really 136 00:07:03,839 --> 00:07:06,200 Speaker 2: focused on, but a lot of that has feels a 137 00:07:06,240 --> 00:07:09,600 Speaker 2: little bit like it's not tangible yet to voters and 138 00:07:09,600 --> 00:07:12,560 Speaker 2: the electorate. What does Rischie Sunak need to do. Number 139 00:07:12,560 --> 00:07:14,280 Speaker 2: one thing he needs to do going into this election. 140 00:07:14,440 --> 00:07:17,080 Speaker 3: So I think there are some really good stories to tell. 141 00:07:17,320 --> 00:07:20,200 Speaker 3: I mentioned t side up in the Northeast when I 142 00:07:20,280 --> 00:07:23,680 Speaker 3: was the Minister for Skills. I was lucky enough to 143 00:07:23,720 --> 00:07:27,320 Speaker 3: be part of this great piece of work where Ben Houchin, 144 00:07:27,360 --> 00:07:30,640 Speaker 3: the mayor, persuaded the energy companies to come and build 145 00:07:30,680 --> 00:07:35,160 Speaker 3: hydrogen plants on the condition that they offered training courses 146 00:07:35,160 --> 00:07:38,600 Speaker 3: for young boys and girls in the local colleges. So 147 00:07:38,680 --> 00:07:40,920 Speaker 3: you have, you know, the government setting up a free 148 00:07:40,920 --> 00:07:43,760 Speaker 3: pot and tea side, the mayor bringing in the business, 149 00:07:44,600 --> 00:07:47,200 Speaker 3: and then the business securing the training places so that 150 00:07:47,280 --> 00:07:50,680 Speaker 3: young people get those new jobs in their community. And 151 00:07:50,720 --> 00:07:53,320 Speaker 3: that is what we call leveling up in the UK. 152 00:07:53,760 --> 00:07:56,800 Speaker 3: And we're starting to see the fruits of it, you know, 153 00:07:57,560 --> 00:08:01,080 Speaker 3: new business in towns which have been left behind, and 154 00:08:01,160 --> 00:08:03,800 Speaker 3: local people getting those jobs and adding to the prosperity 155 00:08:03,840 --> 00:08:05,840 Speaker 3: of the places where they live. That's something that didn't 156 00:08:05,880 --> 00:08:09,120 Speaker 3: happen under thirteen years of labor. It's happening under the 157 00:08:09,120 --> 00:08:11,800 Speaker 3: Conservatives and I've been proud to be part of it. 158 00:08:11,960 --> 00:08:13,200 Speaker 1: Do you want to break some news. Do you know 159 00:08:13,200 --> 00:08:14,120 Speaker 1: when the election will be? 160 00:08:14,600 --> 00:08:16,760 Speaker 3: Well, i'd love to tell you about the primary. 161 00:08:16,760 --> 00:08:19,040 Speaker 1: I don't know yet. I don't know yet. Okay, all right, 162 00:08:19,080 --> 00:08:20,960 Speaker 1: we'll leave it there. Thank you so much for your time. 163 00:08:21,000 --> 00:08:24,440 Speaker 2: That was Alex Burghart, UK Cabinet Officer Minister leaning on 164 00:08:24,600 --> 00:08:26,160 Speaker 2: AI in the public sector.