Loading summary
A
Do you think it is plausible, if we ever have another deep mind, that it could become a global leader while remaining essentially a British business? The narrative being that we're nowhere in AI?
B
The narrative is we're nowhere in AI. We don't have particular strengths that we are only focused on sort of catching up and adoption rather than being at the frontier. We can do both of those things and we can do them really better than pretty much most people in the world. Kids are using AI and they're using it in a way where they're finding increasingly it hard to tell between fact and fiction.
A
Is there blue sky thinking going on in government about that scenario? Hello and welcome to the Rest Is Money with me, Robert Peston. Now, as you know, Steph and I have spent an enormous amount of time on this podcast querying whether this government is preparing the UK properly for this AI, economic, industrial, social revolution. And we have been, let's be clear, doubtful that the government gets it. Now, believe it or not, there is a Minister for Artificial Intelligence. His name is Kanishka Narayan. And I'm delighted to say that Kanishka is joining us today. He's here because he wants to tell the world about this 500 million pound investment fund that the government is creating, a fund for investing in AI. But it won't surprise you that I want to talk to him about a whole raft of other questions, such as, is the government changing the education curriculum appropriately for an AI world? Are they thinking about how they'll raise tax revenues when there are no humans in jobs? How will we pay for schools? How will we pay for welfare? So a ton of stuff to talk to Kanishka Narayan about the Minister for AI. Here's our interview. Kanishka, very good to see you. Now I've got open on my phone sovereignai.gov.uk, which appears to be be an advert for all the great British people who made all the breakthroughs that led us to this, this moment. So for years, underrated Ada Lovelace, Alan Turing, the great genius without whom we wouldn't have today's AI world. We've got obviously Tim Berners Lee, we've, we nod at Denis Asabis, an Alphafold, which is his incredible breakthrough where, you know, he's able to sequence all proteins as a result, you know, DeepMind's astonishing prowess. And this was a number of years ago in, in AI. But I'm struggling, if I'm completely honest with you, to understand from this website what it is because this, this website is sort of advertising this up to 4,500 million pound fund that you're launching for AI doesn't really tell me what it's going to do. So this is your opportunity to fill in the gaps and tell us what is this £500 million fund for?
B
Great. Well, thanks, Robert, and thanks for having me on. Let me justify the website and then I'll tell you about the fund. So the website is effectively driven by this sense that in Britain in particular, I used to spend some time investing in tech in the United States. In the us, the reality is pretty good in technology, but the narrative about technology is just way ahead of reality. And if in fact it pulls the reality up and into the future. And in Britain we have the opposite challenge, which is that the reality is pretty good. We have an amazing history and heritage of all the people you talked about and even the present reality is very good. But the narrative just drops, drags us down and again, again.
A
And so the attention, the narrative being that we're nowhere in AI, the narrative
B
is we're nowhere in AI. We don't have particular strengths, that we are only focused on sort of catching up and adoption rather than being at the frontier. We can do both of those things and we can do them really better than pretty much most people in the world. And so the website was a teaser, as we were thinking about launching, to say, let's build up some excitement about the story of Britain when it comes to AI. And then the specifics of the fund are like, the fund was driven by the sense that in the last wave of technology, we let too many companies start by fantastic entrepreneurs, but ultimately leave when they were thinking about scaling. And we want to avoid that with AI. And so the fund is focused on early stage getting indoors with the companies, 1 to 10 million pounds of investment in some of our best AI companies, but companies where we will then not just use the cash, but every other lever of government public policy, procurement, data, visa support, all of that together to anchor those companies more and more in the uk.
A
Kanishka, there's a ton of other stuff we need to talk about, but let's quickly go to a break.
C
We're proud to say that the rest is money is powered by Octopus Energy. This year. Greg Jackson is back to answer another question. Now, this is something that we see a lot that I wanted to ask you about. When you're starting something new, how do you think about risk and momentum?
A
I think people think entrepreneurs love risk, but I'm not sure they do. I hate it. I never buy individual stocks and shares on the stock market. I don't gamble. I think the thing about an entrepreneur for me is I've got more control. When you're working for a company every day, you're at risk of what that company chooses to do with you. If you're an entrepreneur, actually, you've got far more say in what happens tomorrow and next year than when you're at the mercy of your bosses.
C
Nice one, Greg. Well, thanks to Octopus Energy for powering this episode of the Best Is Money.
A
Study and play come together on a
B
Windows 11 PC and for a limited
A
time, college students get the best of both worlds. Get the unreal college deal.
B
Everything you need to study and play with select Windows 11 PCs. Eligible students get a year of Microsoft 365 Premium and a year of Xbox Game Pass ultimate with a custom color Xbox wireless controller. Learn more@windows.com studentoffer while supplies last ends June 30th terms at aka mscollegepc when
C
you need to build up your team to handle the growing chaos at work, use Indeed Sponsored Jobs. It gives your job post the boost it needs to be seen and helps reach people with the right skills, certifications and more. Spend less time searching and more time actually interviewing candidates who check all your boxes. Listeners of this show will get a $75 sponsored job credit@ Indeed.com podcast. That's Indeed.com podcast. Terms and conditions apply. Need a hiring hero? This is a job for Indeed Sponsored Jobs.
A
All right, so let's actually get into some of the detail of this. I mean, actually, you know, one of the proposals I actually made in a book that I published before the general election, it's called how to Run Britain, was indeed that the government, given that we don't have Silicon Valley, I mean obviously we have venture capital funds in the uk, but we don't have either the risk, high risk taking early stage backers of the scale that they have in California. Perhaps the biggest efficiency of all is we don't have the capability to take small companies global. We don't have the scale up finance. So I was making a very strong recommendation in this book that the government is one of the, in a sense few institutions that can take the kind of risks that are necessary for the kind of returns that you can generate and that government ought to do this. But I'm a bit unclear about how much risk your fund will actually take in practice.
B
Yeah, well, I think the diagnosis is exactly right. I used to invest in Silicon Valley and in the uk the UK ecosystem is about eight times smaller now you could say when you scale gdp, that's not totally out of kilter, but broadly we think we can do much better in this context. In terms of risk specifically, I have to place it in the wider context of all the other parts of capital the government is putting up. We have an earlier stage. The riskiest bets that we're taking and therefore very focused in terms of early innovation are through UKRI and Innovate UK that I know my colleague Patrick Valance has spoken to you about. What this is focused on is saying by the time the British Business bank comes into play with 20 plus billion pounds of capital and then the National Wealth Fund, sometimes we lose companies between the very early stage university context and the scale of capital context. So this fund is very.
A
So this is like a round two or a round three.
B
This is round two, we call it seed or series A in kind of venture language. So this is very early stage still. But you've commercially spun out of the university, you're very focused on the commercial context now. And so seed and series A very early Doorstel get inside play all the other levers of government well, so that the companies have the best chance of scaling. And when they scale, because the government is involved early on, we can help them with scale up capital through our other institutions as well.
A
Just to be clear, would the criteria that this fund would use for determining whether to invest, would they be the same criteria as, I don't know, a Sequoia, for example, one of the biggest firms in the world, rooted in California, or I guess we've had on this program very successful UK venture capitalist Saul Klein. Would you be using the same criteria as he would be using or will the government be, I don't know, looking for different standards and different returns?
B
There are two aspects to the criteria here. One is we want this to be a exceptional performing fund because ultimately we're focused on companies that can move the needle for the GDP of this country. And so you have to make sure that you are taken seriously by the wider ecosystem. If I'm a top tier founder, I think this is a valuable thing rather than just a sort of I can't get capital from Seoul, I can't get capital from Sequoia, so I come to this. So we have to hold very high standards when it comes to portfolio level returns. But the second thing is that we also have to say what is different about taking money from government? And for that reason a criterion is going to be are you a company which is relevant to UK sovereignty? We would define that very Broadly, are you relevant to national security? Are you relevant to ultimately a path which makes this a trillion dollar company and therefore has huge economic implications? Or are you relevant in particular to aspects of life sciences and science which we think are core anchors in sovereign build out in the future as well? So there are some distinct areas we are not looking at. Sort of primarily commercial applications, for example, will probably be less likely to meet the filter because we think there's a very active venture market there, so high quality returns, a great fund. But at the same time really focus on this anchor, which is how can the UK build companies of GDP level impact in this country, which is something
A
that government has historically done, simply filling a gap. So would you expect only to invest in, in a firm that comes to you and says, actually we just can't raise the money? Or is the criterion something slightly different from that, which is we can raise the money? Actually, we would slightly prefer not to take it all from Singapore or America and therefore, you know, can you replace them, as it were? So what's the, what is the criterion
B
for this stage, which is still early on, where we think there's a lot of capital available? The role is the latter. So the role is I can probably go and raise this money from a bunch of other people. But the problem has been that every time we've tried to do that historically, we've had to go to the United States, we've had to go to Singapore, we've had to go to Canada. By having UK investors, including sovereign AI around the table, the chances that this remains a great UK company and is supported by, to be anchored here in the UK, those chances go up.
A
Now, £500 million seems like a lot of money, but I was quite struck. I don't know if you've been following the career of somebody who sort of started out with nemesisabisple, called David Silver, who was with him at Cambridge, with him at DeepMind, you know, it's been really, you know, was, was the leader of the Alphago project, which is again, one of those great AI breakthroughs where, you know, AI, you know, defeated a go, you know, this incredibly complicated game ago. World champion, right? He is setting up a business, he launched it only in November, right. Which is called ineffable thinking. And apparently it's already got a value of 4 billion in terms of the fundraising that it's doing, but it's raising hundreds and hundreds of millions of pounds and all of it's frankly from America, as far as I can see. I mean, the lead investor here is Sequoia. Now you might argue that this is precisely the sort of business that we should be trying to keep British. But if he needs hundreds of millions of pounds, British government hasn't got hundreds of millions of pounds so it is inevitably going to be backed by American venture capitalists. Shouldn't we worry about that?
B
Well, look, I think the main thing I'd say is that there are going to be amazing businesses and David has an exceptional background, is going to build I think a incredible company in the uk. There are going to be businesses which will raise from us venture capital investors partly for scale, partly for expertise. Actually we still have some great investors elsewhere who I think are good targets for UK founders. We struggle.
A
It doesn't really worry you that in this particular instance it's not going to be financed by British money.
B
I want David to create a generation defining company and I want him to do it in the uk. If that means that some of the capital comes from abroad, that's okay. What I do want to make sure is that the UK, both in the private sector and in the public sector is being the best partner to him. So he feels like he can keep the company here and achieve the best success he can. That is why we've been very deeply engaged with David. I hope he'll say to you if you speak to him that he has the sort of support he needs from the UK government and it's my job to make sure that continues so he stays here.
A
Now the other thing on the website we were talking about is the other thing on there is a job description and a way of applying for the job of being the managing partner of this fund. Have you had many applications?
B
We have had a ton of applications. I can't to wait wait to talk to you about the appointment. I think the whole country and the ecosystem is going to be surprised at the quality of applications and the quality of the final appointment. I'm very excited about it and that's I think a latent thing which is that more and more people realize not only is this one of the most defining questions for economy and security, the future of AI, but more and more the best place to have an impact. We've seen this in the AI Security Institute on security questions. The best place to have an impact can be the UK government.
A
Will they feel, do you think that they are running a government institution or that they're running a venture fund?
B
I want the day to day experience to be that of a venture fund. I think this only succeeds if we are the best partners. To founders and to the wider set of investors in the ecosystem. People take the government signal seriously rather than as a sign of poor selection. That said, I want the managing partner to feel that they have all the levers of government at their disposal. And so we want to make sure that from a procurement point of view, there are pots of money, which means that they can call on government as a customer playing a role as well, from an investment point of view, that they're able to engage where national security questions are concerned, to understand what is needed in those contexts as well. There is a huge amount government can do on visas and talent support as well, all of that made available. But the decision making, the culture, the pace has to be top tier venture.
A
And in terms therefore of the culture, would this individual and their team get a profit share?
B
I think the one area where we haven't done that is in the profit context. And that's because, you know, if, of course, it's very competitive at the moment in venture, people are making lots and lots of money. But the thing I felt, I quit venture capital to do politics because I felt it would make a big difference. I think there are a number of people who are now willing to focus on purpose and impact if they feel like they can genuinely have the discretion to have that impact. We have been able to.
A
So you're looking for somebody who has maybe already made quite a lot of money but now wants to do public service.
B
We've had exceptional applicants, really top tier, experienced investors in this country and abroad who want to do this job. I feel very convinced that the offer of purpose with discretion is the right offer.
A
And did you think about, however, a slightly different approach, which would have been to essentially have a blended fund, get some private sector money in, because that arguably might have created more of that risk taking entrepreneurial culture if you had, as I say, private sector money blended with yours and then a team that could perhaps share in the profits.
B
Yes. The big thing I'd say is this is not the only shot we're taking a goal. This is one shot. The British Business bank invests in a number of UK venture capital firms, which is exactly doing what you're describing. The government being a limited partner, taking a stake in those funds, but the funds being private sector funds, the right culture, the right decision making in this context. We felt it was important from an information access point of view and the ability to use all the other levers of government procurement visas data, that it was a bit more internal to government, whilst having the culture and the decision making of venture. That's what we're trying to do with this. But it's not an either or. It's yes and both of those.
A
And I want to sort of move the conversation onto another aspect of your responsibilities as Minister for AI in, you know, in a minute or two. But I suppose if I were applying for this job, probably the first question I would ask you, if you're doing the job interview, is, okay, I have identified this amazingly exciting AI business. If we invest, I don't know, £5 million, can't 100% guarantee we'll make a fortune. But we've kicked the tires, we run the due diligence. Looks fantastic. Do I then have to wait six months for the treasury to sign it off?
B
Yeah, the answer to that is exactly no. The whole point of doing this was to set up something that could work at the pace and with a risk culture of venture. And so in every single aspect of design, including, by the way, you know, it might sound like a small thing, the reason that website looks like it does unlike any other government website was because in every small aspect of design and detail, I want this to sound and feel like a venture firm, which is the top tier firm. And so in terms of pace of decision making, we have agreed governance context, which means that it can now move at the pace that you would expect in venture as well. And we want that to be true of every other aspect of working with sovereign AI.
A
So if a firm comes to you and says, look, we can raise this money within the next three weeks from, as I say, this Singapore fund or this California fund. So we need an answer from you within three weeks. You'll be able to do that?
B
Yes.
A
Okay, well, proof will be in the pudding.
B
Well, I can't wait for the first set of deals to be announced. You'll have seen that this isn't just a kind of promising theory. We've actually already done this.
A
Kanishka the rollout of artificial intelligence. The artificial intelligence revolution is such a big subject. So much more to talk to you about. And in particular, I want to drill down into how the education system needs to change. How what young people are taught in schools has to be transformed.
C
This episode is sponsored by Starling, the bank that helps you organize your money, build great habits, and stay in control of your spending. Now, I talk a lot about financial literacy and its importance in our economy. You know, giving people knowledge, but also confidence with money, too. And Starling reckons anyone can be good with money. And you can start simply by checking your balance daily. It gets easier every day and it takes the mystery out of it and then the real value comes from digging into the details with tools like Starling Spending Intelligence so this is an AI powered search bar within Starling Current accounts where you can ask a question about your spending habits and get an instant answer in app. It helps deepen your knowledge and make more informed money decisions. It's a little like having a financial co pilot in your pocket mapping out exactly where your money goes. Search Starling bank to find out more Good with money starts here did you know about one in three people with plaque psoriasis may also develop psoriatic arthritis which causes joint pain, stiffness and swelling. Does this sound like you listen to what it sounds like to be a million miles away? Trymphy Gusocomab, taken by injection is a prescription medicine for adults with moderate to severe plaque psoriasis who may benefit from taking injections or pills or phototherapy. And for adults with active psoriatic arthritis, serious allergic reactions and increased risk of infections and liver problems may occur. Before treatment, your doctor should check you for infections and tuberculosis. Tell your doctor if you have an infection, flu like symptoms or if you need a vaccine. Imagine being a million miles away. Explore what's possible. Ask your doctor about Tremphya Tap this ad to learn more about Tremphya, including important safety information.
A
Trading at Schwab is now powered by Ameritrade, bringing you an expanding library of education with even more ways to sharpen your trading skills. Access new online courses, insightful webcasts, articles, engaging videos and more, all curated just for traders. Plus guided learning paths with content designed to fit your unique interests. No sifting to find exactly what you need so you can spend your time learning to trade brilliantly. Learn more@schwab.com trading. So you are the Minister for AI. I'm assuming you are the very, very, very first Minister for AI in the history of British government.
B
I think the bundle that I've got of AI and online safety is unique, but we have actually had a couple of ministers before me.
A
Have you? No. Right. Who were your predecessors? I didn't notice.
B
Immediate predecessor. So Farrell Clark did this job before me. I think Camrose did the job as well.
A
Okay, very good, very good. Well, it's a shame you're not a Viscount, but maybe it's a matter of time. But anyway, there are so many big public policy issues around when it comes to AI, so I recently had a conversation for the podcast with Tom Blomfield you may have come across who founded Monzo now doing Venture in California with Y Combinator. You know, it was a very wide ranging conversation, but one of the themes we discussed is the sheer pace now at which AI is reaching a level of competence, which means that in particular a whole astonishing numbers of white collar jobs just at this initial phase look as though they are for humans gonna disappear. And you know, if you look at the pace of change in robotics as well, it is, you know, whether you think it will be absolutely transformative of the economy or just a big shift, you know, either way we are going to, we are living through a massive change in the nature of employment. Now he is one of those people who takes the view that over the next five to 10 years, almost every job that you can think of will be replaced by AI.
B
I think income tax will go away pretty soon, sort of five to six year time horizon. I don't think we will tax human labor. I think we will tax compute. The short answer to that, Robert, is I am thinking about those risks. I am thinking about both the employment context, the implications for a wider economy. We set up a future of Work unit that has now been expanded into a wider AI economics institute that the Chancellor set out in her Mays lecture as well. And I want Britain to be the best in the world at having the quickest understanding both existing signals of what is happening in the labor market and within firms, but also thinking long term about exactly, as you say, transformative AI scenarios and where public policy can go
A
in terms of this is another cultural question. We talk a lot about culture today. To what extent? Because one of the problems, and this is been the reason why we weren't prepared, one of the reasons we weren't prepared for Covid. Ministers tend to focus on today's problems and at a push what are the things we need to achieve before the general election. But ministers and government is incredibly bad at preparing for something where you can't put a precise date on when the shock is actually going to happen. So I suppose the question is you're thinking about this stuff all the time, but is there any mechanism for making sure that all the relevant departments of state and their ministers are thinking about this stuff?
B
Look, I think the first thing I'd say is that there is very clear commitment right from the top, the Prime Minister and the Chancellor, that they believe what I believe, which is that the future of AI is the single most important question for our economy right now in terms of capital investment for jobs in the future, for our national security, both cyber as well as drone and counter Drone defense system and even for our clean energy aspirations. When you think about the fact that to bend the cost curve on new technologies, we need AI to play a big role. So right across the board, whatever your objectives are in this government, those objectives are only served if you take AI deeply seriously as well. There is a genuine question of what do you do now? Given the uncertainty about the timing and the scale to prepare for all contingencies, one of the best things that we can do now, the lesson from every industrial transition of the past is that the countries that are moving fastest on adoption are able to capture the upside of that in the greatest possible way. So that's why we're focused very much on adoption. I'm happy to talk about that in a second. But the second lesson is that there is agency, both for firms, for individuals, for governments, in what type of AI adoption we're doing. There are some firms, and I've spoken to Tom about this, I actually appointed him one of our AI ambassadors very early on in my job. There are some firms that Tom Blomfield is investing in which are looking at total automation and substitution, taking away the human aspect, doing just AI. There are other firms. I've spoken to the CEO of one of the largest tech consulting firms globally who said this is the biggest graduate intake year in their history. There's a robotics listed firm in the UK which has had the biggest headcount this year as a result of adopting AI in a way that supports human labor, grows top line relative to Chinese firms, and as a result is able to become a bigger labor force as well. And so the question for me is how can we move fast in adoption, how can we move fast in augmenting human labour, not just automating it and really at the heart of it, not just in our politics in Whitehall, but across the country, make this debate more nuanced about not pausing AI or speeding up AI, but actually shaping the direction of it as much as we can as well.
A
If you think there is a realistic chance of the so called world of plenty of, you know, where the marginal cost of production is nil because it's all done by basically robots and there are no human employees. And therefore we've got to find a way to fund the welfare state, fund schools, fund hospitals and provide, you know, basic income to people, is there blue sky thinking going on in government about that scenario?
B
So we're in the future working at now, the Economics Institute, we are thinking about long term scenarios very much including that. But the thing I'd say is I personally don't believe in that scenario and here's why I don't believe in it. Whenever you have significant industrial change of that sort, even if you think this is very unique in the scale combining with the speed, which is the rare thing that's happening here, people who make money want to consume things. And when people want to consume things in a overall economy, what ends up happening is that the wages of the things being consumed still in demand go up very materially. You get lots more jobs in those areas as well. So you have this.
A
Well, if those jobs can't be done by robots, yeah.
B
So what I don't believe is that every single possible thing that a human could want could be done by robots forever. I think there is no sort of fundamental cap on the number of human jobs even if you assume very significant advances in artificial intelligence.
A
All right, and then there's the other. You know, you pay money, takes your choice, you know what probability it is. But it is definitely one of those absolutely catastrophic long tail risks which is also, you know, what work is being done in government on what's known as P doom, which is the idea that as and when we get to, you know, serious artificial general intelligence, you know, the all powerful godlike AI will attempt to wipe us out.
B
We are very rare as a government globally in having access to models before they're deployed in the wild to be able to test them for a series of risks.
A
And so including the extinction risk.
B
Well, I think the extinction risk plays through a series of different things happening. One aspect of that, of the sort of predom scenario is loss of control. This is models being able to effectively run off and do things without human understanding or human engagement. And we test models for a series of risks, cyber, biochemical, in some cases, loss of control. And we do that not just with a view on the catastrophic extinction stuff, but with a view on saying what does this mean for the governance and risks embedded in our critical national infrastructure private sector today? And so we are very unique in understanding those risks, understanding them before others and therefore being able to act on them as well. It happens to be an input into the possible Ptoom scenario as well. So of course it informs my view and keeps me worried about where things might go. But for now, the most critical aspect of that scenario is a scenario that also affects things today.
A
And as it happens, I've just finished writing a new thriller called Kill Switch which is all about whether we'll be in a position to shut down a rogue AI.
B
So we talk about this through the Lens of alignment. We have a whole fund focused on this. We're doing a lot of work to make sure that models are very much aligned with what we're trying to do as humans and as a society.
A
So now let's sort of talk about the potential benefits to earlier adoption. So you've got a strategy, you say, to help roll out productive, useful AI in public sector and private sector.
B
That's right. So on the private sector side, the adoption strategy is effectively all the horizontal things, all the things that apply to all kinds of sectors and firms. So skills, we're training up 10 million people across this country in AI skills by the end of the decade.
A
What does that mean? Actually, I've seen that pledge, but what does it mean in practice? How are you training up 10 million people?
B
Yeah, so that specific program, rather than the broader skills aspect, that specific program is focused on two things. One, we are partnering with the largest employers in the private sector and public sector. So the NHS is a part of that. But we also have large firms, Accenture, for example, we have firms like Cognizant as well. We have a series of large firms that we're partnering with. And we're saying, can we together design training that will most quickly move the pace on adoption in your firms and with the talent that is coming into your firms in the future as well. And so we're working together with them to make sure that those companies are scaling up the workforces that they're engaged with. Separate to that, we have a program where we're effectively allowing everyone in the country to be able to come online and for free through a government funded program, be able to learn the starting building blocks of AI usage in their sectors as well. And so there are specific kind of basic level things there. There are slightly higher level courses as well. We're driving the uptake of that through a series of other partners as well. So both specific places that we think can move the needle, but also a general program.
A
And what about schools? Because one of the, I mean, we had Mark Warner, a faculty also on the podcast, and you know, one of the things we got into a bit of a loving over the fact that it's completely mad. It seems to me that we, the current sort of rule in schools is you don't use AI. I mean, literally my head wants to explode when I think about it. Given that even, you know, if you're right, that what we're really talking about is AI more in terms of certainly in the next few years augmenting our productivity rather than Replacing us, obviously it's going to replace quite a lot of jobs. Then obviously young people have to be comfortable using it right from the word go. So have you talked to Bridget Phillipson, the Education Secretary, about this issue that broadly at the moment, schools ban it?
B
Well, we're speaking, Robert, on a day where actually I'm on my way later to launch a series of technical excellence colleges focused on digital technology and AI. And so we have spoken and we have also started to act very significantly on it.
A
But does she get it?
B
Yeah, I think she does, actually. I think Brigitte is very focused on how we can equip our kids for the future. And the thing I obsess about the most is I think all of government and frankly, all of our society should be focused on one question, which is the question I obsess about. How can we spend every minute of our time making sure that kids growing up in this country have jobs in the future which are fulfilling? And also the cognitive context which allows them to operate in a world of AI? What I mean by that is the context I'm seeing loads and lots is whether schools are formally doing it or not, kids are using AI and they're using it in a way where they're finding increasingly it hard to tell between fact and fiction. And so it's on us to make sure that we're using our curriculum, using our teaching context, to equip kids with a much deeper understanding of what is true, what is not true. And that cognitive ability to be able to manage AI is ultimately a really core skill that we need to build.
A
Yeah, but I mean, there seems to me to be two other aspects of all of this. I mean, so the example, which I literally couldn't believe from a particular school was, you know, a bunch of kids were being taught coding. And as you know, you know, whether it's, you know, anthropics, cowork or, you know, whatever. Broadly, vibe coding is now a thing. Natural language coding is replacing an enormous amount of the coding industry, and yet these kids were told they couldn't use AI in the classroom. And this is just literally insane.
B
As you know, Robert, you know, Claude code, codecs, cursor, all these tools allow you to automate a bunch of lower level coding. What is still really important is software engineers who understand fundamental architecture can then manage and oversee the coding agent. And so we need to build those skills. We can't build those skills if we're still telling our kids that they have to do everything manually. We have to teach them how to use the calculator when they're doing maths, rather than just do it entirely in their heads?
A
Totally.
C
Right.
A
And then I suppose the other aspect of the school's challenge is we do have to think very radically now about what are the areas of the jobs market where we think for the foreseeable future, AI won't be as good as humans. Or even if AI could possibly be a substitute, that actually is, you know, consumers won't want AI. They will want real humans. And then think about, you know, therefore, what are the skills within schools that, you know, kids should focus on? And they probably aren't. Well, the curriculum is definitely not the curriculum that would be ideal for that kind of AI world. And again, you know, do you. Are you having these con. These curriculum conversations with the Education Department and Bridget Phillipson, or are they doing it autonomously, as it were, without you needing to be nudged?
B
I think there are two parts to this. One is the curriculum conversation. The second is how do you design a system where the education system and the job matching system moves dynamically as the economy changes? So on the first of those, we're looking at a series of specific sectors. I'm actually launching a campaign very soon focused on chip design. We have far two kids in this country who currently think about chips and semiconductors. A huge part of the economy of the AI economy in the future, and also a part which is very hard to automate, given the extreme precision that you need in some parts of it and the understanding you need of different parts of engineering. And so there are some areas where we're making specific bets. And then the second question to me is, as a country which has one of the most centralized information on who has come to a job center? I remember going to Cardiff Job center and giving my CV at the age of 16. It's probably in a drawer somewhere. We have tons and tons of this text data, interview data from all of us who've gone to the job center, and then an understanding of what has happened when someone gets into a job. There is no better place to support the dignity of people looking for a job and deploying frontier technology in the service of that dignity than in better matching jobs to where the economy is moving. And if we can do that radically better, that is another way in which we don't just have to get the bet right in terms of what's going to be the next job, but the system adapts itself and more quickly than when the economy moves.
A
Finally, I just wanted to focus on some of the sort of more microeconomic short term challenges. So obviously I've seen that OpenAI is launching a research center in London which I think potentially will create quite, you know, a few hundred important research jobs in, in London alongside DeepMind. And you know, London is now, as you know, you've been saying very much. Well, the UK is very much ahead of the, the rest of Europe when it comes to digital and AI development and London in particular is looking pretty healthy. But OpenAI did announce a few months ago, big fanfare from the government, Stargate uk, which was this massive investment in a data center and compute and has shelved it. And they've shelved it because energy's too expensive and because they were also unhappy about the data regime and their access to British data. Tell me first of all, how much of a blow was that when OpenAI decided to shelve that?
B
Look, I want companies to invest in computing data centers. It's a pretty important part of infrastructure. The main thing I'd say on OpenAI is that when the commitment was made, since then there has been no change to energy pricing and there has been no change to the regulatory environment. What has changed is, as well publicized is OpenAI's internal financing context has become much more challenging and they have made similar, similar calls in the United States with their program there as well. So you know, companies get into trouble, companies have challenges in financing commitments they've made. I don't want to hold anyone to kind of ransom on that front. The thing I'd say broadly, but everybody
A
does say let's be absolutely. Well, I mean not only do they say it, it's objectively true. It is objectively true that the cost of energy in the UK is higher than in almost any of our important competitors. I mean that is a competitive disadvantage.
B
Sure. I'm not at all disputing that. Our energy pricing can be radically more supportive of AI in the future. It is obviously more expensive. That is just a factual claim. What I would say is that at the moment we already have a good market share in both Europe and globally in that I want it to be even further. And so what we're doing across policy is a huge.
A
So you think it matters to have sovereign data compute capacity? Because some would say we've just lost that race.
B
I don't think it's a sort of either or so the United States has 100 plus gigawatts of compute capacity, 75% of global share.
A
And you know, Stargate in the, in the US is apparently still going ahead and it's, you know, $500 billion?
B
Yeah, I think, I think it's going ahead with a smaller scale, so that's fine. I don't think it's the right thing for the UK to aspire to anywhere near that level of development. I do think it's really critical for us to have core parts of compute in this country that are important for health, financial services, defense areas where we need data here in the country, protect our ability protected and our ability to allocate that compute to ends that are strategically important is really critical.
A
How long will it take to get the relevant capacity? Because it feels painfully slow.
B
Well, the one thing I'd say is I think under the radar we've made a series of reforms that have effectively ring fenced AI growth zones and data centers from a lot of the micro challenges you're describing. AI growth zones will now be first in the grid queue. From an order point of view, we will put money into local authorities so those local authorities accelerate planning. We will offer them a concierge service and a connections acceleration so that they're able to move fast in the grid context. If they build in areas like Scotland, the northeast and Cumbria, where you have curtailment, excess generation, the savings from them building there will get shared back in part with the investors. So we've juiced up the economics of it. So right across the board, from a planning point of view, a grid point of view and a grid economics point of view, we're making it as attractive as possible to make sure that companies are building the data centers that we need in this country.
A
I suppose in your heart, do you think it is plausible that if we ever have another deep mind here, that it could become a global leader while remaining essentially a British business? Because the one thing that we don't have here, which is obviously an enormous advantage to any business that basically has what economists would call network qualities, is the UK market is just tiny compared to the American market. And one of the reasons why America has been so unbelievably competitive as an economy and so unbelievably good at creating world leading businesses over the last century, it's just because the domestic market is huge, right? We're not in the European Union single market anymore. When we were, when we as a nation were the leading force in creating the single market, it was precisely because we understood the economic benefits of market scale. But we don't have access to that anymore. So I suppose as I say, it takes back to my question, absent at some point in the future, rejoining the European Union single market. And we have a Prime Minister who says he wants sort of little by little to have more single market access, but he's not talking about full access. Could you be confident that a deep mind would achieve what Hassabis has now achieved if it remained British?
B
I think about this question all the time and let me just say two things on it. One, I don't think the question of rejoining the eu. And by the way, I want us to have a very productive relationship with the eu. And the Chancellor and Prime Minister have been very clear on that. I don't think that is the single most pivotal question in whether or not we get a future DeepMind. And the existence proof is that DeepMind existed prior to Brexit. And DeepMind obviously found a owner in Google who was better able to finance its aspirations. And so whether or not a trade relationship expands your market enough to keep your biggest shots and goal like that,
A
certainly the financing infrastructure in their case was incredibly important.
B
Google invested something like 5 to 7 billion pounds over a 10 year period. And Google was able to do that because it was a huge company with network effects and a return on invested capital of 25% year on year. We don't have that here.
A
People like me, I've spent my entire professional life wishing for Britain to have that kind of world leader.
B
And my single biggest aspiration in this job is for the UK to create a trillion dollar company with all those effects, to create amazing jobs here so that never again in the future do we have entrepreneurs here who don't have the financing possibility from a national champion in the way that I hope the sovereign AI Fund will help contribute to as well.
A
All right, well look, on that optimistic note, I'm sure our listeners wish you luck. There are big challenges and it's an incredibly interesting time to be doing what you're doing. But, but Kanishka, as always, very good to see you. And that's it for this episode of the Rest is Money. Goodbye from me.
B
Why did we really go to war with Iraq? And did Saddam Hussein really have weapons of mass destruction? I'm Gordon Carrera, national security journalist. And I'm David McCloskey, author and former CIA analyst. We are the hosts of the Rest Is Classified. And in our latest series, we are telling the true story of one of history's biggest intelligence failures, Iraq WMD. In 2003, the US and UK told the world that Saddam Hussein had weapons of mass destruction.
A
But they were wrong.
B
This wasn't a simple lie. It was something far more complicated, far more interesting. And far more dangerous spies who believed their sources, politicians who wanted the public to believe in the threat, and a dictator who couldn't prove he'd already destroyed the weapons. In this series, we go deep inside the CIA and MI6, go into the rooms where decisions were made, and look at the sources who fabricated the intelligence that took us to war. The Iraq war reshaped the Middle east and permanently weakened part. Public trust in governments and intelligence agencies and its consequences are still playing out today. Plus, in a Declassified Club exclusive, we are joined by three people who were at the heart of the decision to go to war. Former head of MI6 Richard Dearlove, Tony Blair's former communications director, Alistair Campbell, and former acting head of the CIA, Michael Morell. So get the full story by listening to the rest is classified and subscribing to the Declassified Club. Wherever you get your podcasts.
Date: April 15, 2026
Hosts: Robert Peston, Steph McGovern
Guest: Kanishka Narayan (UK Minister for Artificial Intelligence)
This episode centers on the UK government’s ambitions to remain a major player in global artificial intelligence (AI) innovation and economic transformation. Hosts Robert Peston and Steph McGovern interview Kanishka Narayan, the UK’s Minister for Artificial Intelligence, about the newly launched £500 million sovereign AI investment fund, Britain’s unique strengths and challenges in AI, and the policy shifts needed for the UK to build, retain, and scale world-leading tech companies. The conversation ranges from venture capital culture and risk-taking, through the disruption of jobs and the education system, to existential risks and the UK’s infrastructure for AI.
[00:01–05:00; 07:01–08:57]
"The narrative is we're nowhere in AI... But the reality is pretty good. We have an amazing history and even the present reality is very good. But the narrative just drags us down..."
— Kanishka Narayan [03:31]
[05:00–11:59]
"So this is like a round two or a round three... This is round two, we call it seed or series A... but you've commercially spun out of the university, you're very focused on the commercial context now."
— Kanishka Narayan [08:57]
"Are you a company relevant to UK sovereignty... national security... or aspects of life sciences and science which we think are core anchors in sovereign build out?"
— Kanishka Narayan [10:02]
[11:59–14:47; 39:16–46:21]
"If that means that some of the capital comes from abroad, that's OK. What I do want to make sure is that the UK, both private and public sector, is being the best partner to him so that he feels he can keep the company here..."
— Kanishka Narayan [14:19]
"We don't have that here... My single biggest aspiration in this job is for the UK to create a trillion dollar company with all those effects..."
— Kanishka Narayan [45:40], [46:02]
[15:04–18:19]
"We've had exceptional applicants, really top tier, experienced investors in this country and abroad who want to do this job. I feel... the offer of purpose with discretion is the right offer."
— Kanishka Narayan [17:04]
“The answer to that is exactly no. The whole point of doing this was to set up something that could work at the pace and with a risk culture of venture.”
— Kanishka Narayan [19:12]
[23:26–29:41]
“Over the next five to 10 years, almost every job that you can think of will be replaced by AI.”
— Robert Peston, summarizing Tom Blomfield [24:40]
“I think income tax will go away pretty soon, sort of five to six year time horizon. I don’t think we will tax human labor. I think we will tax compute.”
— Kanishka Narayan [25:12]
“I don’t believe every single possible thing that a human could want could be done by robots forever... There is no sort of fundamental cap on the number of human jobs even if you assume very significant advances in artificial intelligence.”
— Kanishka Narayan [29:44]
[29:57–31:55]
“We are very rare as a government globally in having access to models before they're deployed to be able to test them for a series of risks.”
— Kanishka Narayan [30:29]
"We have a whole fund focused on this...to make sure that models are very much aligned with what we’re trying to do as humans and as a society."
— Kanishka Narayan [31:45]
[32:15–39:16]
"We are partnering with the largest employers...can we together design training that will most quickly move the pace on adoption in your firms and with the talent..."
— Kanishka Narayan [32:35]
"Kids are using AI and…are finding it hard to tell between fact and fiction. So it’s on us to make sure that we’re using our curriculum, using teaching, to equip kids with a much deeper understanding of what is true, what is not true."
— Kanishka Narayan [34:50]
[39:16–43:15]
“Our energy pricing can be radically more supportive. It is obviously more expensive. That is just a factual claim. But at the moment, we already have a good market share.”
— Kanishka Narayan [41:21]
“Core parts of compute in this country...important for health, financial services, defense areas...to allocate that compute to ends that are strategically important.”
— Kanishka Narayan [41:48]
[43:15–46:21]
"The UK market is just tiny compared to the American market. And we don’t have access to [the EU] anymore."
— Robert Peston [43:15]
"I don't think EU membership is the pivotal question...DeepMind found an owner in Google who could finance its aspirations."
— Kanishka Narayan [45:00]
“My single biggest aspiration...is for the UK to create a trillion dollar company with all those effects, to create amazing jobs here so that never again...do we have entrepreneurs here who don't have the financing possibility from a national champion in the way that I hope the sovereign AI Fund will help contribute to as well.”
— Kanishka Narayan [46:02]
On narrative and reality:
"The narrative is we're nowhere in AI...reality just drags us down and again, again."
— Kanishka Narayan [04:06]
On fund purpose:
"The fund was driven by the sense that in the last wave of technology, we let too many companies...ultimately leave when they were thinking about scaling. We want to avoid that with AI."
— Kanishka Narayan [04:54]
On government’s competitive edge:
"People take the government signal seriously rather than as a sign of poor selection."
— Kanishka Narayan [15:43]
On taxation and future of work:
"I think income tax will go away pretty soon...I don’t think we will tax human labor. I think we will tax compute."
— Kanishka Narayan [25:12]
On existential risk:
"We are very rare as a government globally in having access to models before they’re deployed in the wild to be able to test them for a series of risks."
— Kanishka Narayan [30:29]
On education reform:
"We can’t build those skills if we’re still telling our kids that they have to do everything manually. We have to teach them how to use the calculator when they’re doing maths, rather than just do it entirely in their heads."
— Kanishka Narayan [36:54]
This episode is a rich, candid look at how the UK government is striving to shape its own AI future, balancing optimism, realism, and policy innovation to build the companies, talent, and infrastructure needed for the next great economic leap. If you want to understand not just the hype but the high-stakes decisions and strategies behind Britain’s tech ambitions, this conversation is essential.