Latest / AI in the UK / Fast-Tracked Power for Data Centres? Chancellor Launches UK AI Sovereign Fund; Fighting Fraud in the AI Age
Transcript
- 0:00All right, your luck. Listen up.
- 0:15Hello and welcome to AI in the UK.
- 0:18I'm Neil Footman. And I'm Matt Wilkinson.
- 0:21Professionally, we help organisations, leaders and
- 0:24individuals learn about and implement this big scary
- 0:27behemoth called AI. There's some amazing stuff going
- 0:31on with AI in the UK that doesn't always get the attention
- 0:35it deserves. With AI in the UK, we want to
- 0:38change that. Hi and welcome to this week's
- 0:57edition of AI in the UK. On this week's show, the UK
- 1:01government wants to fast track National Grid connections for AI
- 1:05data centres. Sensible strategic move or a
- 1:08flagrant misuse of public resources?
- 1:10Can Rachel Reeves new Sovereign AI fund keep UK AI companies
- 1:17British? And with online fraud hitting
- 1:19record levels, do we need an entirely new paradigm in the age
- 1:23of AI? But before we move on to our
- 1:26stories, Matt, I gather you were at a very interesting event this
- 1:30week with one of the major AI players.
- 1:33Yes, so I was fortunate enough to be part of the Association
- 1:36for Key Account Management Annual Forum.
- 1:39And this year Oracle, we're hosting it.
- 1:42So we were in their centre in the middle of London.
- 1:44As we've talked about, you know, London is a huge AI hub and, and
- 1:48Oracle are becoming one of the biggest players in building
- 1:51that, you know, AI infrastructure.
- 1:53And their ambition is to really help companies connect what
- 1:58Oracle are calling Islands of opportunity.
- 2:00So for for them in particular, you know, Oracle's been, you
- 2:04know, a pretty acquisitive company over the years.
- 2:06So they've built up what many companies might call silos or
- 2:11islands of knowledge. And So what they're trying to do
- 2:14is to build, connect the different buckets of data using
- 2:19AI agents and really being able to unlock the power of that
- 2:22data. So what they're building for
- 2:23themselves, they're then making available for customers.
- 2:27They currently have about 1000 agents, you know, embedded
- 2:31across their Fusion apps, which are kind of some of their really
- 2:33big data suites. And I know that they've got
- 2:36plans to release multiple agents in this quarter alone.
- 2:39And so that pace isn't incremental.
- 2:41And so they're really trying to look at how can they, how can
- 2:44people at all levels of an organisation with data and
- 2:48insight rather than just, you know, reports, you know, trying
- 2:51to suggest what could that next great step be?
- 2:54Really, really interesting and exciting stuff.
- 2:56I see. And what capacity were you there
- 2:59in? So, so I'm on the board of the
- 3:01Association of Key Account Management and gave a small,
- 3:05small talk on how AI can be used and some some examples of that
- 3:09for sort of key account managers and building sort of dashboards
- 3:12and sort of agents and that that kind of fun stuff.
- 3:14But it was a day full of everything from storytelling to,
- 3:18you know, how to influence so a so a real smorgasbord of
- 3:22interesting insight. Fascinating.
- 3:25And by happy coincidence, the whole infrastructure theme leads
- 3:29nicely into our first story this week, which I'll pass over to
- 3:32you for. Yes.
- 3:33So the UK government is consulting on plans to fast
- 3:38track National Grid connections for AI data centres in it's
- 3:42designated AI growth zones. You know, that's in response to
- 3:46a 400 160% surge in grid connection requests in just six
- 3:51months last year, with speculative applications
- 3:55clogging the queue and leaving viable projects facing waits of
- 3:58up to 15 years. The proposals would give
- 4:01priority access to AI data centres and growth zones, while
- 4:05stricter entry conditions aim to flush out projects with no
- 4:08realistic prospects of being built.
- 4:10The consultation on the plans closes in April.
- 4:13The announcement coincides with a sharply contested planning
- 4:17decision in North Lincolnshire. On the 11th of March, the
- 4:20council unanimously approved the Elsham Tech Park, a proposed One
- 4:25GW AI data centres campus near Scunthorpe.
- 4:28According to the council's own planning documents, the
- 4:30facility's peak annual electricity emissions would
- 4:33exceed 1,000,000 tonnes of carbon dioxide equivalent by the
- 4:37early Twenty 30s. Campaign group Foxglove argued
- 4:41their developers figures understated the carbon impact
- 4:44fivefold and that the council overrode its own policy
- 4:47requiring on site renewable generation.
- 4:50Together, the two developments expose a growing tension at the
- 4:53heart of the government's AI ambitions.
- 4:55The infrastructure required to deliver them is already testing
- 4:58the limits of both the National Grid and the country's climate
- 5:01commitments. So Neil's what do you?
- 5:04What do you make of this it? Covers a lot of ground here and
- 5:07so it touches on a lot of issues that do come up again and again
- 5:10regarding data centres, which are a massive growth theory and,
- 5:14you know, inherently controversial.
- 5:16This tension I think between central strategy and local
- 5:21accountability is one that's going to become increasingly
- 5:24contentious as more of these things are built out.
- 5:27I think in general I am a fan of devolved power.
- 5:31I think local areas are better equipped to make decisions
- 5:36affecting the people that live there.
- 5:38I understand this is a complicated trade off and AI is
- 5:41a massive deal. I would question.
- 5:44It seems to me in certain in certain respects this is kind of
- 5:47how it's supposed to work in that planning accountability is
- 5:52supposed to be held at a local level.
- 5:54I think the idea that you could just centralised this and move
- 5:58these decisions entirely to to Whitehall does not equal better
- 6:02or more efficient decisions. But particularly when there's so
- 6:07much emphasis, every government comes in and talks about
- 6:09levelling up the regions and giving the regions more more
- 6:12autonomy and more financial power.
- 6:14So surely this is a good opportunity to put the money
- 6:18where the mouths are. That's really hard to to argue
- 6:21against. I think my big push back though
- 6:24is the North Lincolnshire Planning Committee has approved
- 6:27a project that predicts that it's going to emit the
- 6:30equivalent of every UK domestic flight annually and those,
- 6:34that's the number that they've used to approve the project.
- 6:37Foxglove have said those factors, you know those figures
- 6:41are wrong by a factor of 5. So is that a local planning
- 6:45decision if we're going to all of a sudden be emitting from one
- 6:48side 5 times the UK domestic flight sort of amount of carbon
- 6:54dioxide kind of makes it puts to get, you know, puts in
- 6:57perspective just the impact of these data centres and how we
- 7:02should be looking at. It no doubt.
- 7:03I mean I don't think you can reduce this to a simple central
- 7:08is better versus region is better.
- 7:10This is in any branch of government.
- 7:11This is a constant source of tension.
- 7:13So can central government be sufficiently accountable versus
- 7:16can local government be sufficiently strategic?
- 7:19I don't know. I don't know the perfect balance
- 7:21and clearly they don't yet either.
- 7:23It's something that will be a source of debate and contention
- 7:27as long as these things have been rolled out, I would
- 7:29imagine. And it kind of feeds into, I
- 7:32mean on the broader point, the idea of the government giving
- 7:37priority to giving priority access to the National Grid for
- 7:41data centres. I think that so with big
- 7:46caveats, this is this is not simple either, but I think there
- 7:49is actually quite a lot to be said for that if you merely.
- 7:54So the policy as it has been so far, broadly speaking first come
- 7:58first served. The difficulty with that, as
- 8:01we've seen is it has kind of fed an atmosphere of speculative,
- 8:06speculative acquisition of power resources for data centres that
- 8:10may or may not take place. The idea underlying this, this
- 8:15plan of having a more strategic outlook and giving priority
- 8:18access to to real credible data centre developments, I think can
- 8:23only help in in can only help in apportioning and controlling
- 8:29power use. First come, first serve, when
- 8:31there is likely to be such a massive surge and there's
- 8:34already such a massive surge in demand for a specific type of
- 8:38energy consumption, I think is not going to work with with
- 8:42massive caveats about accountability and so on.
- 8:45It seems to me a central government strategic approach
- 8:47could very well be what's required here.
- 8:49It does. Sound like a sound strategy, but
- 8:51it also feels a bit like a subsidy.
- 8:53You know, we're handing scarce public infrastructure to private
- 8:56operators based on a growth narrative.
- 8:59You know, there's no published criteria for how those priority
- 9:02decisions get made. And I think there's also the
- 9:05other challenge that we've seen where some of those commitments,
- 9:09as you alluded to just now are commitments.
- 9:11There's nothing actually legally binding for them to man, you
- 9:14know, to actually build these these data centres.
- 9:17It's just essentially A commitment.
- 9:19And so it's a letter of intent as much as anything else.
- 9:22So there's. So that's that's so it's
- 9:24interrupts. So that's that's your
- 9:25understanding. So I'm, I'm not gainsaying, I'm
- 9:28your understanding is that it's no more than a commitment, but
- 9:31it's, it's a stronger commitment than a kind of speculative,
- 9:35speculative approach, but but no more than that.
- 9:37It's a kind of as you put a kind of memorandum of understanding
- 9:40rather than any kind of contractual obligation to
- 9:43actually follow through and build these things.
- 9:44Yeah. And I guess the other question
- 9:47is, is who's auditing the list and actually at what point do
- 9:50you know at what point in the the sort of the planning process
- 9:54of these connections being made? Because actually until you've
- 9:56got a connection request, until you know that you're going to be
- 9:59able to connect, you're not going to start start building
- 10:01the data centre. So it's AI.
- 10:03Completely agree that it's a chicken and egg scenario, but it
- 10:07does feel to me like like it could be a little bit of pink,
- 10:10let's put it that way. OK, well the other big point
- 10:13perhaps that this story raises is around the emissions, which
- 10:15are, on the face of it, pretty concerning, to say the least.
- 10:19Shocking might be a more more suitable description.
- 10:22It seems to me, though, that a lot of these predictions are
- 10:25just that they're projections, and as we know, projections are
- 10:28famously always wrong. These are scope 2 emissions,
- 10:32meaning that they are not direct emissions from a power plant or
- 10:37emissions generated as a result of using power generated by
- 10:41power plant, is my understanding of what what that means.
- 10:43So if if the government targets are met over the next few years,
- 10:48I think it's reasonable to assume that the output from
- 10:50somewhere like an Elsham would be considerably lower than
- 10:54current projections, say, which changes the calculus of the
- 10:57entire project. Judging it on today's priorities
- 11:01and on today's standards and any any kind of these projections as
- 11:04well, they'll be coloured by current political considerations
- 11:07and the current energy market and so on.
- 11:09It seems perhaps a bit short sighted.
- 11:11Yeah, I think it's it feels a little bit like an accounting
- 11:14trick in some ways. You know, they're drawing, you
- 11:16know, these projects are drawing from the grid at scale without
- 11:19actually adding to the power generation capacity itself.
- 11:22And we know we already have some of the most expensive
- 11:25electricity in the world. And of course, with current
- 11:28geopolitical tensions, you know, that's pushing up the way that
- 11:32our, you know, our, our prices are calculated for electricity.
- 11:35So I do believe that operators do need to not only be looking
- 11:39at the the capacity that they want to build in terms of what
- 11:42they're building, but they should.
- 11:43There needs to be some kind of requirement to fund some, some
- 11:47kind of new capacity to keep the, you know, to keep the
- 11:50lights on really not just for the data centres themselves, but
- 11:53also to make sure that we're not putting the grid under too much
- 11:56stroke. So you mean, sorry, you mean the
- 12:01operators of the data centres should have to commit to
- 12:03producing additional capacity themselves, I think?
- 12:06There should definitely be some kind of contribution there.
- 12:10Maybe part of the planning should be that the data centres
- 12:13need to cooperate with the power generation companies to actually
- 12:17be able to to make sure that their power needs a map.
- 12:21Don't know but might be making things overly complicated.
- 12:23But it just feels that if we just keep bolting on more and
- 12:26more demand, we could end up with a situation with sort of
- 12:28rolling brownouts and not enough power to keep the lights on.
- 12:32I mean, it's not, it's not completely beyond the realms of
- 12:34possibility that you do hear about in the US, some of the
- 12:38hyperscalers funding the construction of light nuclear,
- 12:42small smaller nuclear reactors to power some of these data
- 12:44centres. Quite how that would play in the
- 12:47UK. I'm not sure if Microsoft wanted
- 12:49to open a nuclear power plant to support data centres, but in
- 12:53principle you'd think it's it's not not such an outrageous
- 12:56suggestion, yeah. It's one that I think we should
- 12:59definitely be at least considering as we as we look at
- 13:02sort of what it what what's currently happening with our
- 13:04energy prices. Yes, absolutely.
- 13:08OK. And moving on to our second
- 13:10story this week, also closely concerning a government
- 13:14initiative. So in her right, OK, I should
- 13:17have checked this beforehand. Maze lecture.
- 13:19Is that what was pronounced Mais?
- 13:22As far as I know, yes. OK, so let's go with that in her
- 13:24May's lecture in London this week, Chancellor Rachel Reeves
- 13:28made a direct pitch to British A A companies stay in the UK and
- 13:32the government will back you. The sovereign AI unit is a £500
- 13:36million unit launching on 16th of April.
- 13:38And the pitch goes beyond money. The headline offer is compute
- 13:42access through Isimbard AI, EU KS most powerful AI
- 13:45supercomputer launched in Bristol last July.
- 13:48But for commercial company training a frontier model, the
- 13:51more meaning thankful commitments may be quieter ones.
- 13:54The 1st is capital. The unit works alongside the
- 13:56British Business Bank and Innovate UK to invest directly
- 14:00in high potential AI startups, backing companies at the point
- 14:04where many are most vulnerable to acquisition by larger, often
- 14:07American, rivals, £500 million is a serious commitment at seed
- 14:12and early stage, but it's less clear what happens when a
- 14:14British AI company needs a billion pounds or more to train
- 14:17its next model. The second is procurement.
- 14:20Reeves was explicit. She wants government contracts
- 14:22to become a launchpad for British AI companies, rather
- 14:25than, in her words, a check for global incumbents.
- 14:29Recurring public sector revenue let's the company build without
- 14:32surrendering equity or control to offshore investors.
- 14:36The third is talent. The unit is expanding the ENCODE
- 14:40Fellowship, designed to attract the world's leading AI
- 14:42researchers into British labs. It is run by a private VC Pillar
- 14:47VC, in partnership with Aria, so not purely a government
- 14:50initiative. But there's a harder question
- 14:53underneath all of this. For many venture backed
- 14:55founders, acquisition by Google, Microsoft or Meta is not the
- 14:58threat, it is the plan. The government is asking British
- 15:01entrepreneurs to choose a more difficult path over a
- 15:04potentially lucrative American exit.
- 15:06Whether that ask is realistic and whether procured contracts
- 15:09and fellowship programmes are sufficient to change those
- 15:12incentives may be the most important question the unit
- 15:16faces. I thought this was a fascinating
- 15:18story for several reasons. There's there's obviously a
- 15:20whole lot going on there and it's as with many of these
- 15:23things, the devil will be in the detail.
- 15:26So there's a couple of points I'd raise about this map before
- 15:29kicking over to use Stop me. If you've heard this one before.
- 15:31The scale of money being talked about South £500 million.
- 15:36It is when you hear again and again about funding rounds for
- 15:40AI giants that are reaching 10s or hundreds of billions, it's
- 15:43kind of puts it in perspective. That said, if it's, you know, at
- 15:48seed stage, it is a significant amount of money when we get to
- 15:51much bigger scale up size, it isn't.
- 15:54And this speech specifically did not really push that to the
- 15:57fore. The, the answer, as we've
- 16:00discussed here potentially is to get that kind of money available
- 16:05to, to invest. It really requires a reform of
- 16:08how pension funds invest their money.
- 16:10And she didn't say much about it here.
- 16:12The other point I'd raise on this, I think there was good
- 16:15stuff in there, this idea of persuading founders to keep AI
- 16:21companies British. Do you, do you remember we
- 16:23talked about a company called Faculty that had been bought by
- 16:26Accenture. It was a few weeks ago.
- 16:28Yeah. So that was completed this week
- 16:30and I actually heard an interview with the CEO of
- 16:35faculty who will become the CTO of Accenture.
- 16:39And for him, of course, perhaps he's not going to say it was all
- 16:42about the money. So the money of course is a big
- 16:44factor. But what was absolutely critical
- 16:46for him was the exposure that being bought by one of the
- 16:49world's biggest consultancies could afford.
- 16:51Even if a pension fund in the UK could raise that kind of money
- 16:54to to invest, could they give them that kind of exposure to
- 16:57all the world's blue chip companies?
- 16:59And that was as much of an incentive as the money itself.
- 17:03And you hear something similar actually from the story of Deep
- 17:05Mind. As we've discussed many times,
- 17:07Dennis Hassabis does not come across as someone who's
- 17:10primarily driven by money. The real attraction of the
- 17:13Google deal for him was the ability to work with the
- 17:17biggest, most well resourced labs in the world.
- 17:20And it really subsequently did really power Deepmind's
- 17:25subsequent research and growth. Anyway, I've talked long enough
- 17:28over to you. What?
- 17:29What do you make of all this? Yeah.
- 17:30So I think this is pretty exciting for, you know, as you
- 17:33say, seed and early stage companies where, you know, that
- 17:37foreign acquisition, you know, pressure is, is greatest
- 17:40potentially. And, and 500,000,000 lbs
- 17:43deployed at that stage really can determine whether a company
- 17:45scales here, here or sells very, very early.
- 17:48You know, the, the fund itself, it, it, it's already funded a
- 17:52company called Open Bind. And you know, that's looking,
- 17:56you know, and that's to the tune of an £8,000,000 investment to
- 17:59create drug binding data sets that are 20 times larger than
- 18:03anything previously produced, you know, anchoring AI driven
- 18:06drug discovery in the UK, at least up to that stage.
- 18:09But I, you know, while and while I welcomed the procurement
- 18:12commitment, you know, I think that that is a durable
- 18:16retention, you know, tool. It doesn't require a company to
- 18:19to sort of stop a company to seek US venture capital in its
- 18:23next funding round. And so that recurring government
- 18:26contracts good. They help keep, you know, they
- 18:28help organisations and keep them going, but it's not going to
- 18:32deliver the scale of entry into, you know, into other
- 18:37organisations that, you know, that being acquired by the likes
- 18:41of an Accenture, as you just just mentioned would give.
- 18:44And I've seen that before where organisations working within the
- 18:47NHS do very, very well with their NHS contracts.
- 18:51But then scaling outside of, you know, that walled garden of, you
- 18:55know, UK, UK government procurement becomes incredibly
- 18:58difficult. And so that's, that's my worry
- 19:01really. And even having funding to to
- 19:03get world leading AI researchers into British labs, does that
- 19:07really mean that we're going to keep AI companies here or just
- 19:11that we're going to do what we so often do, which is build AI
- 19:14or build companies to a scale where we've got some great
- 19:17innovation and then they get gobbled up by, you know, bigger
- 19:20companies on the world scale? Yeah, all the distant points.
- 19:23I mean, I think to to repeat much of what she said is steps
- 19:27in the right direction and, and, and that applies across quite a
- 19:31few of the the government's AI initiatives.
- 19:33There are just big honking problems that it's so difficult
- 19:37to address. I think soaps so much of the
- 19:40infrastructure is promising. Got the, you know, the academic
- 19:42backup. We've got the the tax incentives
- 19:45now we've got some some more funding for early stage
- 19:47companies until we find a way to provide access to the types of
- 19:53very large amounts of capital required.
- 19:56Once these companies start to scale, then this is going to
- 19:59keep happening. And even if, as I said, even if
- 20:02you can secure that level of funding from within the UK,
- 20:05you've still got the problem of something like what?
- 20:09Mark Warner, he's called the the the CEO of faculty that some of
- 20:13these enormous industry leading companies offer advantages that
- 20:17just money can't. That's the access they can
- 20:20provide to the exact type of customers that that that they
- 20:23want to deal with. And just as an aside on on the
- 20:27whole scale point, the supercomputer is an apparently
- 20:32seriously impressive piece of kit.
- 20:35But it's again, when it's taken in comparison to competitors in
- 20:40the likes of the US, it is in terms in terms of not just
- 20:44capability, but access. How many people can actually use
- 20:47this thing is it is, I understand prioritise for
- 20:49researchers, which is all well and good.
- 20:51Can it really support big industrial ambitions?
- 20:54I'm not so sure. So to repeat, to repeat a
- 20:58recurring theme throughout our stories, steps in the right
- 21:01direction looks great. Have to address the the access
- 21:04to bigger amounts of capital and that surely requires pension
- 21:08reforms or, or investment pension fund investment reform,
- 21:11yeah. I'd agree it's one of those
- 21:13challenges where we're looking at AI in the context of the UK,
- 21:18and the UK does have some structural challenges on the
- 21:20global stage and AI alone won't fix those challenges.
- 21:24Indeed. OK, so from from the Mace
- 21:28lecture from Rachel Reeves, we'll move on to our final story
- 21:30of the day, which addresses something that's an age-old
- 21:34problem with technology, but in something of a new light.
- 21:37Matt, what's going on here? Yes, Cephas, a not-for-profit
- 21:41fraud prevention service, has published its annual fraudscape
- 21:45report this week and this. The report covers data from 2025
- 21:49and it shows that the National Fraud Database recorded 444,000
- 21:55cases of fraud, which which was a 6% rise on 2024.
- 21:59So you may think, oh, it's only 6% up, but that's still nearly
- 22:03half a billion cases of fraud and, but it and it's the highest
- 22:07annual total ever. And before we think it's just
- 22:10fraud, it's probably credit card or whatever.
- 22:12Actually 54% of all cases were identity fraud and you know that
- 22:19did fall slightly year on year as criminal shifted tactics
- 22:22toward account takeover, which now represents 18% of all
- 22:26filings. The sharpest misuse was misuse
- 22:28of facility, up 43% and Cephas members prevented an estimated
- 22:342.4 billion in losses. But the gap between fraud volume
- 22:38and enforcement response continues to widen.
- 22:41The organisation frames the underlying driver as AI enabled
- 22:45industrialization, synthetic identities, deep fake
- 22:49impersonation and fraud as a service toolkit that allow low
- 22:53skill operators to run high volume attacks.
- 22:56Criminal networks are now being structured like corporations,
- 22:59specialised teams for data gathering, scam execution,
- 23:02payment processing and laundering.
- 23:04Over 1200 cases were recorded every single day in 2025.
- 23:11That's a pretty frightening number really.
- 23:13The structural problem is that the defences remain largely
- 23:16individual and reactive. Bank alerts, victim awareness
- 23:19campaigns, identity verification at the point of onboarding.
- 23:23While the attack surface is systemic and automated, the CFS
- 23:26report anticipates AI will be used to build long term victim
- 23:30profiles and personalised attacks.
- 23:33Further, there's no real coordinated government response
- 23:36to AI enabled consumer fraud currently visible.
- 23:39And so that becomes a a bit of a warning, you know, with that
- 23:43barrier to, you know, to running 10,000 simultaneous
- 23:47impersonation attacks, you know, we can do that, you know, just
- 23:50from a laptop and a cheap API subscription.
- 23:53We're not just dealing with more fraud, we're dealing with a
- 23:55categorically different threat. And that threat will typically
- 23:59be targeting some of those that are, you know, the most
- 24:02vulnerable in society. And we really need enforcement
- 24:05frameworks that are fit for, you know, fit for purpose.
- 24:10And the ones that are currently there that are built for human
- 24:13scale crime are underfunded and pretty much obsolete.
- 24:16What's your take? Niels, my initial take is there
- 24:19are certain phrases that emerge from time to time to make you
- 24:22despair for humanity and fraud as a service is definitely one
- 24:26of those fast. Remember when you heard it
- 24:29first, folks? Did you come across that before?
- 24:31Fraud as a service? I haven't, no, but it's it's
- 24:34scary, isn't? It, it is and it's, it's
- 24:36entirely believable that how well organised and resourced a
- 24:39lot of these operations will be. So look, I find it very
- 24:45difficult to disagree with the fundamental idea that the
- 24:48current frameworks we have in place are probably not suited.
- 24:51I do think yes, we could have been using them better
- 24:55previously. I had a quick look and just to
- 24:57see how the UK compares in international standards for
- 25:00prosecuting fraud. And we're not particularly good
- 25:03and we're not, but we're not that bad either.
- 25:05Apparently the, the the picture internationally is, is pretty
- 25:08poor, which would suggest perhaps need for greater
- 25:11international cooperation on these things because obviously
- 25:13frequently the people actually perpetrating these crimes aren't
- 25:16nowhere near the UK that, you know that they're operating out
- 25:19out of countries far away. I think though, perhaps what's a
- 25:24pertinent issue and something to keep in mind is the solution
- 25:29worse than you know, is the cure worse than the disease?
- 25:32And I think it's something that you definitely have to watch out
- 25:34for in this regard. So underlying I think what, what
- 25:39you're suggesting Matt is that there would be greater
- 25:41government coordination and and a sort of more central
- 25:44coordinated role for the authorities in dealing with this
- 25:48in general. Yes, I think we're too far the
- 25:50other way, but be careful what you wish for.
- 25:53I would say so, actually. So the one country that ranked
- 25:56very well for prosecuting fraud is Singapore.
- 26:00Now people have quite disparate views on Singapore.
- 26:02I've only been once. Have you been to Singapore?
- 26:04Only transitly, right? So.
- 26:06What you would say about Singapore or what is frequently
- 26:08said is it's a beautiful place, everything works, is the
- 26:12corruption is very low. Government generally functions
- 26:15and that they pay government people very well.
- 26:18But for some people, the degree of control over day-to-day lives
- 26:22is something that they're just not comfortable with.
- 26:24And there's, I'm sure there's far worse examples, far more
- 26:26authoritative regimes than Singapore's.
- 26:29And some people may elect to go to Singapore that they're
- 26:32willing to make that trade off. They will sacrifice a bit of a
- 26:34bit of personal freedom because they prefer, you know, an
- 26:38orderly and law abiding and functioning society.
- 26:41This is a choice to the individual.
- 26:43But there's every chance, I would say, if you try to entrust
- 26:48the supervision of online financial transactions to that
- 26:52degree, you are in danger of moving towards a much more
- 26:56authoritative approach than you would perhaps be comfortable
- 26:59with in other circumstances. Yeah, I'd I'd agree.
- 27:02I think the challenge the the challenge is obviously getting
- 27:05the balance right. But when you look at who is most
- 27:11vulnerable to to these fraudulent attacks, it's clear
- 27:15you can't just run awareness campaigns to get your, you know,
- 27:18to get yourself to a point where we become more resilient than
- 27:23others and push the, you know, push the defraud to other
- 27:26countries. And that wouldn't work anyway
- 27:28because people are targeting individuals rather than nation
- 27:31states. And you really can't sort of run
- 27:34awareness campaigns that that are going to work where
- 27:36criminals are structured like corporations running these
- 27:39automated attacks. You know, if they're being
- 27:41successful 1200 cases a day, how many are they?
- 27:44How many attacks are they actually making?
- 27:46I mean, it's, it's quite frightening.
- 27:48You know, the response has to match that attack surface.
- 27:50Something coordinated needs to be done.
- 27:53But you're right. How much is that over
- 27:55regulation? And how much is that towards us
- 27:58and the use of our tools versus, you know, versus sort of
- 28:03actually going after enforcement?
- 28:05And how do you go through enforcement if that fraud is
- 28:08happening in organisations? Well, well, in countries where,
- 28:12you know, the long arm of the law can't reach it's that's a
- 28:15real, real challenge here. And you know, just telling
- 28:18people to spot phishing emails. It's that's not a policy.
- 28:21And being careful as to how you answer the phone and these as
- 28:24these attacks get so, so sophisticated and the profiles
- 28:28they can build about us improve, it's quite scary.
- 28:31I mean, I now longer answer any question, if somebody phones me
- 28:34up and you know, from any of my suppliers, I don't trust that it
- 28:38that's them from the suppliers I have to phone up.
- 28:40And that, that lack of trust that I now have of anybody
- 28:43contacting me directly is, is quite scary.
- 28:46You know, can you log in Well, well, actually, you know, what
- 28:49is the safe word or the, the password that the organisations
- 28:52have to give us to make sure that we know that they're
- 28:55genuine There, There's some, there's some big questions
- 28:58there. Absolutely.
- 28:59Actually I looked up the numbers again prior to, prior to us
- 29:04getting together this week and just one, one indicative stat
- 29:08said there were around 4000 convictions last year for online
- 29:12fraud in the UK against 1.23 million reported offences.
- 29:17And I think that just there could barely be a starker
- 29:21demonstration of just what enormous difficulty this is and
- 29:25why current frameworks are not likely to work.
- 29:28One other point before packing. But also there is there is
- 29:31another risk to having a very centralised oversight of fraud
- 29:35intelligence. The more, the more you
- 29:37centralised anything, the more you centralised any kind of
- 29:39online or, or, or, you know, electronic oversight or database
- 29:43or whatever it is, the more attractive that database becomes
- 29:46as a target for hackers. So it's not as if, so I agree we
- 29:53need to move in that direction, but it's not as if that approach
- 29:56is risk free in itself, both in terms of personal freedoms and
- 29:59in terms of fraud becoming a target for fraud itself.
- 30:02It is a desperately complicated, but I would add desperately
- 30:06urgent issue that we need to deal with as a society.
- 30:10Great. I mean the I mean you're a
- 30:12director of a business like me and having to go through that
- 30:15one Gov digital identification process to get your code.
- 30:19That frightens me, the amount that the government is actually
- 30:22storing about me in a single system.
- 30:24You know, that, that I had a few problems with the app.
- 30:26I think it was being overwhelmed at the time.
- 30:28So I had to go through their question sets and actually the
- 30:31questions about my, my personal finances that it was, it was
- 30:35frightening exactly how much they seem to know about me.
- 30:38And, and that's when I haven't even approved who, you know,
- 30:40proven, proven who I am. That's just part of trying to
- 30:43prove who I am. And so I think it's kind of
- 30:45scary having all of that data in one, one place about me.
- 30:49That data set would be a, you know, surely has vulnerabilities
- 30:53and if anybody gets into that, that, you know, just think about
- 30:57somebody having all of the details from every single
- 31:00company director in the country. That's a frightening thought in
- 31:02itself. It doesn't bear thinking about.
- 31:05It really doesn't. So let's not on that.
- 31:09Yeah, on that note, it's always good to end in a positive,
- 31:14positive. It hasn't happened yet.
- 31:15No, it hasn't happened yet. So my, my, my defence is going
- 31:18to be to maintain a very boring existence.
- 31:21So even if they do get on my personal details, they'll find
- 31:24nothing of value there. That's going to be my.
- 31:26That's going. To be my strategy, right?
- 31:29With that, that brings us to the end of this week's show.
- 31:32Thanks as ever Matt, for your wisdom and insights.
- 31:36Thank you for yours as well, Neil.
- 31:37It's been great to talk. Yeah, and we'll see you next
- 31:39week. Bye for now.
- 31:40Bye. Bye in the UK.
- 31:51And that's it for this week's episode of AI in the UK.
- 31:56If you found it useful, you can subscribe wherever you get your
- 31:59podcasts and we'd really appreciate a great rating or
- 32:03review. It genuinely helps others find
- 32:05the show. You can also follow Neil's
- 32:08Footman and Matt Wilkinson on LinkedIn for links, notes and
- 32:12occasional extra content. Thanks for listening and we'll
- 32:17be back next week. Helping you make sense of AI as
- 32:20it lands here in the UK.