430. Murdoch, Epstein, and AI Billionaires: Is Trump Losing Control?

Thanks for listening to The Rest is Politics.

Sign up to The Rest is Politics Plus to enjoy ad-free listening.

Receive a weekly newsletter, join our members' chat room, and gain early access to live show tickets.

Just go to the restispolitics.com.

That's the restispolitics.com.

The rest is politics is officially powered by our friends at Fuse Energy.

And something else we all know is that energy deals they're about as stable, long-term, and reliable as a, I don't know, Donald Trump promise to a a country.

But not this one.

Fuse Energy is offering cheap, fixed tariffs for up to 18 months.

No price hikes, no drama.

Just a rock-solid rate you can count on.

And with the July 1st price cap, now's the perfect time to lock in a better deal.

And the average home is going to save about £100 compared to the price cap.

And that's not spin, that is real money back in your pocket.

And here's the clincher.

When you sign up, you get da-da-trip for free.

And you know by now, that means afree listening, it means bonus content like our recent mini-series on JD Vance, early access to question time, and pre-sale tickets for our sell-out live shows, powering your home and your brain at the same time.

Head to fuseenergy.com/slash politics, download the app, and use the code Politics to get a free Trip Plus subscription.

Fixed tariffs, big savings, bonus content, what's not to love.

Welcome to the Rest is Politics.

Me, Alistair Campbell.

And with me, Rory Stewart.

And I'm in London.

Rory's in Columbia.

We should talk about Latin America when you get back, Rory, because we don't do enough on Latin America.

But today, I think we should talk about,

here we go again, Donald Trump, this time in the context of this pretty extraordinary story about Jeffrey Epstein.

You want to talk about AI, and I've been, since you said you want to talk about AI, I've been digging into it and I've ended up more alarmed than when I started.

And I want to celebrate what you will call the lowering of the voting age and what I call the democratic extension of the franchise.

So let's celebrate.

One of, finally, one of my campaigns comes off.

Votes at 16.

I've been banging on about that one for ages.

Anyway, what do you make of Donald Trump and Jeffrey Epstein?

Well, yes, let's start with that.

Let's try a Rory disagreeing agreeably with you.

Okay.

And this may be because I'm in Columbia, therefore quite a long way from the center of the story.

My assumption, and this may be completely naive, is that Donald Trump is a man who had a lot of mistresses, a lot of it just to show off, loved taking women to nightclubs, was very much parading as a kind of 1980s, 1990s cool guy.

He's obsessed with looks.

You you've often pointed out that he's often praising male looks.

He loves, you know, his cabinet being sort of matinee idol people.

That's what he at least believes about Pete Hegseth and the rest of them.

But I'm very doubtful that there's any evidence at all that the guy is a pedophile.

And my assumption is that he hung around with Jeffrey Epstein because Jeffrey Epstein was a kind of fixture of New York society and Trump liked powerful, wealthy people.

And that therefore, you know, I hate Donald Trump and I want him to disappear.

But I imagine probably the thing that he's not guilty of and all the other things that he's guilty of is sleeping with underage minors.

But anyway, over to you.

Well, I assume from your introduction, saying we're going to disagree agreeably, that you think I'm going around the place saying Donald Trump is a paedophile.

I am not.

But it is sort of pretty remarkable that we're now in a state of affairs where you can start the UK's number one political podcast by saying, I don't have any evidence that Donald Trump is a paedophile.

What we do have evidence of is that he was very, very close to Jeffrey Epstein, that he is, as you say, obsessed with the idea that when you're famous, you can pretty much treat women in any which way that you want.

And also, that the Epstein story goes to the heart of what could become, I don't think we should overstate it, but I think this could become a problem for him within the MAGA movement.

Because as we all know in life, sometimes the biter gets bit.

And in this case, I would say if you live by the conspiracy theory, there is a chance you might die by the conspiracy theory.

Because this is now a story that, I mean, we don't get that much of it in the UK, but there's a whole kind of ecosystem in the US, particularly on the right, about Jeffrey Epstein.

And it started because their beloved conspiracy theory, and this is part of what drove Q and On, who thought that the Democrats were running the world, and they were basically a bunch of satanic, blood-sucking paedophiles working out of a pizza parlour.

I mean, the thing is so mad, it is incredible that it's become part of the body politic.

But essentially, they believed that the Democrats were, and particularly the Clintons, were at the heart of all this and that Epstein was the key to it.

He kind of got all his money through blackmailing these people because they were all in it with him.

And of course, Trump and the MAGA crowd, they fueled these conspiracy theories.

And then when they got into power, You had that Pam Bondi, the Attorney General, saying, the Epstein files are on my desk.

It's sitting on my desk right now to review.

That's been a directive by President Trump.

She said she was going to publish them.

They published something that was next to nothing.

And then she came out and said, actually, there's not much in there at all, thereby fueling the conspiracy even further.

And it just feels like one of those stories that is never, ever, ever going to go away.

And that's what happens with conspiracy theories.

Fried Zachariah loves this joke that

somebody ends up at St.

Peter's gate at heaven and he says to St.

Peter, hi.

And St.

Peter says, you have one question.

We will answer one question.

He says, Okay, who killed John F.

Kennedy?

And St.

Peter says, Lee Harvey Oswald.

And the man says, Oh my goodness, the conspiracy goes much deeper than I thought.

God is in on it.

But

this will now fuel all sorts of new conspiracies.

There's the thing about Ghillaine Maxwell, who I knew very well because, of course, I worked with her dad, Robert Maxwell.

And she actually, I don't know if I've told you this, Roy, she introduced me to Geoffrey Epstein the one time I've met him.

It's the reason I'm in the famous black book.

Very hurt that he spelled my name wrong.

And Alistair, to tempt, I mean, just sorry, that's quite a big revelation.

I don't know many people.

Sorry, I know you and some very, very famous people who met him.

And most of the famous people I know who met him

don't talk very much about him.

They don't pretend they did.

Give us a sense of your memory.

of him and and above all of Ghillane, who clearly I mean, these are clearly people who did really unpleasant, profound wrong.

And this thing I do know a little bit about, so just a second introduction, people who aren't really focusing on this, they set out deliberately to recruit very young women, many of them underage from very vulnerable backgrounds, and exploited them for sex over many, many years, leaving some of them profoundly traumatized and damaged, including in certain cases, them leading to them taking their own lives.

Yeah, absolutely.

And they used all their wealth to cover that up.

And their political access and their connections, including famously with a member of the British royal family.

I mean,

I knew Ghillaine pretty well when Maxwell ran the mirror.

And she was very energetic, very lively, quite funny.

Always got the feeling she was his favourite child.

And I sort of socialised with her a bit, but then completely lost touch.

And then I can't even remember exactly when it was, but I was still in government, still working for the government.

So it was more than 20 years ago.

And I was on a plane going to a funeral in America.

And I was just sitting there minding my business.

And she was, she also happened to be on the plane.

She was sitting a few rows behind me, came up, had a chat, and then just said, and I can, I can definitely remember this: her saying, What are you doing tonight?

And I said, Not much, I'm just going to be staying in the hotel, etc., etc.

And she said, Do you want to come and meet my boyfriend?

And so, yeah, well, nothing to do.

I always quite liked Gillette, I must be honest.

I thought she was a very, you know, really warm personality.

I always felt a bit sorry for her because her dad was such a monster.

So I went to the place that Prince Andrew was famously photographed leaving.

Thankfully, I was not photographed either arriving or leaving, therefore, no innuendo could be laid at my door through the photograph.

And when you say this place, this was his fancy house in New York.

Yeah, yeah, and it was very fancy.

It was very fancy.

It was incredibly luxurious.

It reminded me of, you know, those sort of really old-fashioned sort of university-style clubs where I'm sure you've got them at Yale, Yale, you know, where old alumni gather to sort of, you know, talk about the old days.

It felt quite old-fashioned, but you know, a lot of very fancy art.

He sat for a while behind this kind of, it reminded me a little bit of the Resolute Desk, a really big desk.

He had photographs of himself with famous people all over the place.

I didn't see if there was one of Trump.

I mean, Trump was probably not remotely on my radar, but definitely of most of the presidents of our lifetime.

He was, he was an incredible name-dropper.

I got a bad vibe, and not because of any sense of him, of all the sexual stuff that we since know about, but just the name-dropping and the kind of arrogance, and there's something a bit sly about his face as well.

Ghilaine left us alone, just the two of us, and then she came back about an hour later, and I said I was tired and wanted to leave.

So, that was it.

That was my experience of him.

Can I ask a question?

Because I guess this is to the heart of a lot of these scandals, which is psychologically what it feels like when somebody who you've met, who you've liked, who you've known for a long time, Ghillain Maxwell would be an example, is then revealed as having done truly horrendous things.

And what happens psychologically, does one initially feel a little bit defensive towards them?

Sorry, let me be more personal before I put you on the spot.

I've been in the experience where people that I know have got themselves involved in the most horrendous scandals.

You can think of politicians that we both know, priests, teachers.

And I often find there's a moment of dissonance where I think, no, no, wait a second, I know this person, I've known them for many years, I can't really imagine them doing this, and therefore, that moment of wanting to come to their defense because you feel that they're being slandered, and then it turning out that actually this thing is true and that they've done something absolutely horrendous.

Did you go through that process with her, or have you gone through that process with anyone else?

I mean, how does it?

I mean, I don't, because it's sort of so, that was the last time I saw her.

So that was a couple of decades ago, at least a couple of decades ago.

I think a part of me, of course, don't forget, I also knew the sons, I knew Kevin and Ian in particular, and you know, they've been through horrendous scandals and the stuff about the pensions and then their father's death and all that stuff.

So

I think my

when I first heard about the Epstein stuff, which had been bubbling away, but when it first became sort of something that was front page rather than kind of inside and was leading the news, I can't say I was surprised that he was involved in real scandal because he struck me as a bit of a sleazeball.

I had a sense of Ghillaine as being.

I was surprised.

I was surprised.

I can imagine that children who've had the sort of parent that they had,

I can imagine could end up with all sorts of sort of psychological stuff going on.

Do you know the thing I feel at the moment?

This is why I think she's probably now going to try and have another go.

She's going to claim that she's had an unfair trial because she thought she was covered by this deal in a previous era.

I do find it incredible that she's the only one who's gone to jail.

Was I surprised that he ended up in this sort of scandal?

I mean, the scale of it, yes, but the fact of it, probably no.

I think when you look at those videos today of him and Trump and the way they were behaving together, and that's when there's a camera on them, remember.

Just imagine what it's like when there is no camera on them and their attitudes to women.

I don't think that saying when I knew her that Ghulane was a warm, quite attractive personality is defending her.

It's saying what my experience of her was.

I don't remotely and never would defend what she has done in league with Epstein.

But you know, I do think though that this has the capacity to do Trump quite a lot of damage.

It was interesting the other night.

He was on Truth Social.

You know,

Michael Walforbis says that he can't, when he's worried about something, he can't sleep.

And he was sort of posting absolutely insane stuff through the night.

Videos of women and snakes, and you know, yes, a few attacks on his usual people, but stuff about football and stuff about baseball.

And I mean, mad stuff.

And I just wonder if he is feeling a little bit more defensive about this than he lets on and I guess the the other thing to add to this is that Elon Musk has now broken with him very very dramatically and Musk is completely obsessed with this issue in fact Musk sees himself as a crusader to protect children against sexual predators and that often involves him very very madly projecting this.

You remember there was this case of trying to save Thai children who'd got stuck underground.

Yeah, yeah, yeah, yeah.

And him accusing the man man who was trying to save them of being a paedophile.

A lot of his campaigns against other figures before he broke with Trump were about the Epstein files.

And of course, Musk is now talking about setting up a third party and he hopes that he can spend, let's say, 500 million pounds buying three or four senators, getting a few congressmen on side, and then having the balancing vote.

in the next election in Parliament, where he holds the balance of power in Congress and his key number of senators or congresspeople can actually determine which way a vote goes, which would give him enormous power.

So, I think in terms of real politics, this question of whether Elon Musk is going to be able to fan this conspiracy theory, drag some of the MAGA base over to his side or not.

And let me just make the argument against, then I'll come back to you.

My instinct still is that Musk doesn't really understand politics and that he's not a very popular figure.

And that actually, as Zoran Mamdani showed in the New York race, you can't actually buy politics in the the end of the US.

That in the end, Zoran Mamdani beats someone who had 80 times as much money as he did, Andrew Cuomo.

And my suspicion is that Elon Musk will fail.

But that is another dimension to this, whether Musk is going to be able to use his obsession with the Epstein files and the MAGA base's obsession with the Epstein files to be able to make this move.

And to underline, it is an obsession.

These people are utterly obsessed with this stuff.

And people like Marjorie Taylor Greene, who would, you know, in normal circumstances, be seen as sort of diehard Trump, but she is a conspiracy theorist and she believes all this stuff.

And she feels cheated.

That the Trump and Co., having said we're going to publish all this stuff, the files are going to be out there, you're going to see how the Democrats are at the heart of this, she feels cheated.

And of course, the other big figure on the right of American politics, who's now right at the center of it,

getting on in years, even older than Trump, of course, is Rupert Murdoch.

Because the Wall Street Journal, edited by my old friend Emma Tucker of the Financial Times, and they ran this story.

I mean I think in normal circumstances this might have been back in the day when Trump and Epstein were real sort of chums about town.

It would have been a kind of page six of the you know diary story the fact that he was sort of sending bawdy messages.

Now he's the president and we know what we know about Epstein and we assume that that was going on back then as well.

This is like a big deal and of course you've now got Murdoch and apparently Trump and senior people from the White House literally virtually got down on their knees to Murdoch and said, please don't run this story.

And Murdoch said, no, paper's going to run it.

And they run it.

And he's now suing them.

And of course, he's had some success extorting money in this way.

So this thing's going to become the latest chapter of the reality TV show.

But he seems less able to control the script on this one.

I'm now going to make an enormous improbable lurch in our coverage by going from the sort of reality TV show.

to the existential threat.

And I'm going to justify it by this, that one of the problems, and this is because I'm going to go on and talk about some of the challenges and threats posed by the latest artificial intelligence models and what might happen over the next two, three years.

And this is genuinely probably potentially one of the biggest risks that humanity has ever faced.

And the reason I'm making this shift is that we are going through this development at a time of Trump, at a time of populism, at a time when, in fact,

unlike in the post-war era where when nuclear bombs were developed, when you you had a stable bipolar world, Soviet Union against the US, development of the United Nations, development of NATO, multilateral institutions to control it, AI and its threats are coming in a world which is disintegrating, where there really is not a sense of grown-ups in the room and where it's very difficult to believe we're going to be able to get international codes and values to control it.

So let me come over to you on that.

You've been reading a bit about AI.

What struck you in reading about it from a policy point of view?

I think the thing that struck me most is that when you go and read, well, I know you've been seeing this guy, Yoshua Bengio, and I've been reading stuff from Geoffrey Hinton, who was one of the first to sort of say the kind of thing that you've just been saying, is really their sense of alarm at the speed at which this is developing and their fear that it's not just that governments won't be able to control it, but humanity won't be able to control it.

And I read a paper that that they both put their name to alongside lots of other experts, and

it was really quite profoundly scary, basically saying that it won't be long, like, you know, within years, when these AI systems are going to outperform us as human beings.

And on the one hand, they're saying this could help our living standards, it could protect our environment, it could alleviate poverty in the poorest parts of the world.

But that alongside that, there are genuine societal and risks to social stability,

risks to our understanding of what the world is.

And I think we're seeing that already just with the sort of social media revolution.

And the way that I think about it, and of course, at the moment, you're in Columbia, you probably won't be aware that today, one of the big stories on the news is that the British government has done a deal with one of the big AI firms to help them deliver progress in schools, in the justice system, in the healthcare system, and so forth, which, which, you know, it's part of their strategy for growth, it's part of their strategy for trying to be part of this AI revolution, which is going to be so important to the economy.

But I sort of, I remember when social media first came along thinking this is really exciting, it's going to break down the old media oligarchies.

But in fact, we've ended up with a worse situation with just a new breed of media oligarch who is now into this AI world as well.

And I sort of feel, I feel that about this: I don't, I can't see how you get the national and international governance around this stuff that makes it a force for good so I on the back of you saying you wanted to talk about it and the back of me going and reading a few books and papers I have become a lot more pessimistic if I can be allowed to do a very brief explainer for people who aren't in the world of AI all the time and actually I'm slightly worried by the fact that often when we put out things on AI less excited than I would expect given that this is I think the technology which will probably change the world more in the next three to ten years than anything else conceivable.

It's quite interesting that people are a bit reluctant to hear about it.

But here's a brief explainer just to remind people where we are.

So an incredible thing happened through something called the Transformer architecture.

And there's no reason to go into the technology of it, but there was a sudden exponential jump in the capability of machine learning.

And anybody now using the latest models from ChatGPT or Gemini or even Claude or even the Chinese version DeepSeek, and who knows how to prompt them, will find incredible advances.

And there's been a second advance relatively recently, which is in their ability to reason.

So if you use the O3 models, you'll often find, for example, that they will say, thinking, okay, the user asked me this, but I need to think about that.

That is suddenly meaning that we have had an exponential improvement just over the last few months in their ability to do maths, their ability to do science, their ability to do computer programming.

And as you say, the benefits of this are extraordinary, particularly, for example, in medical science, where AIs very, very rapidly are going to make extraordinary improvements, very straightforward ones, so things like radiology picking up on tumors, but pretty soon in their ability to create new medicines.

However, you will also have picked up when you are using these AIs that they can get things wrong.

And it's difficult to to talk about machines in a human way, but they behave in a way that with a human, we would call deceptive.

They are not entirely honest with us.

And that comes from two problems.

Firstly, they're trained on an internet, which is full of human lies.

And secondly, in the post-training phase, they're often encouraged to please us by bluffing, partly because the programmers reckon that we don't really like AIs saying, oh, I don't really know the answer to that question, I'm not sure.

So we encourage them to be certain.

And there are real problems with this.

And one example that people talk about a great deal is an AI trying to access a website found itself confronted with capture.

You remember that thing which you often get on a website?

Tick this to show that you're a human.

The AI then reached out to a task grabbed human and said, Would you please tick this for me?

The human said, How do I know you're not a bot?

And you can see in the thinking, the AI saying, well, you know, if I say I'm a bot, it probably won't agree to do this, so I better pretend to be a human.

So the AI says, I'm visually impaired.

The human ticks the thing for it and it's able to access.

So it's being deceptive, it's being manipulative.

Now, why is this a problem?

Then I'll stop talking.

Well,

because, firstly, in defense, we can see increasing pressure in Ukraine to develop autonomous systems.

In other words, because there's a problem of drones being jammed by the Russians, for example, there is pressure to develop systems which no longer need to talk to their controllers at base.

They just fly off and they make their own decisions on the front lines very, very rapidly.

There will be a huge problem in politics.

That isn't just deep fakes.

That's also the fact that these machines can persuade you.

They can create QAnon, but QAnon which is very, very localized, very much individualized for you.

a particular you know Alistair or Rory version of QAnon and they can groom you over weeks and months to become more and more extreme.

And finally, economic concentration of power.

If ChatGBT eventually developed an advanced form of artificial general intelligence, it is of course theoretically possible for OpenAI to say, we're not going to release that model to the world.

Instead, we're going to use it to set up all our own companies, which will become the most successful companies in the world.

And then we will suck all the revenue of that back to ourselves.

So we become the biggest companies in the world.

And other countries that don't have these things, their tax revenue base will collapse.

So that the defense, political and economic power will then reside in these companies.

And that's why I think we need to think about how we make this a global public good.

How do we think about how, instead of all this power being in the hands of five or six companies in the Chinese government, the power of this and the benefits of this and the security control of this can be vested in something more global.

Yeah, well, that's the other comparison that I make with the earlier stages of the tech revolution is that in theory, you could be talking about something that is

profoundly equalizing, where we all have access to the same information, we all have access to the same learning tools, we can all make money in the same way.

It doesn't matter where you came from, doesn't matter what country, we've all got access to this.

But I think I need to pray and aid again the sovereign individual, because this was all laid out in Mr.

Reese Mogg Sr.'s plans for the world, that this is about the rich and powerful becoming more powerful because they know how to do this stuff.

I mean the thing I find, you know, when even when you're talking there in fairly simple, straightforward terms,

I struggled sometimes to work out what this actually means in terms of how it works.

And so therefore that sense of this being a driver of increasing inequality with the rich becoming richer, the powerful becoming more powerful, and the powerful not including necessarily the politicians.

And that's why I think, look, what you said there about essentially, and when we were talking about this earlier in the week, you were saying that what worries you is the extent to which this is being pitted as a battle for the future between America and China, right?

So they are the two big players in this.

So countries like Britain necessarily feel we've got to get involved, we've got to get engaged in this.

But I think we're inevitably coming at it from a place where we're giving even more power, even more influence to people who aren't necessarily good actors.

Yes.

There are two points that the tech companies make at the moment.

One of them is you can't regulate us because if you do you will slow us up and we need to get there before China.

If we can get there, they will argue two years before China, then our superintelligence, our AI systems will be able to control the Chinese systems.

That's the first argument they make.

The second argument they make is you can't regulate us because you don't really understand what you're regulating yet.

This technology is developing so fast that any regulations you put in place will be out of date immediately.

It would be like trying to the early days of regulating cars where we tried to approach them as horses.

There's something in that, of course, but the risk is that you're putting an enormous amount of trust in private companies.

I mean, they're effectively saying, we're the good guys.

Please don't get in the way, government.

Trust us and our own internal safety procedures to handle this stuff for us.

Please don't insert government officials into the process because if you do, they won't understand what they're doing.

They'll just slow us up and China will win.

Why would we trust Musk and Zuckerberg and these guys?

What have they done other than make themselves lots of money by being very clever?

What have they done for the world?

Why would we trust them?

So Alistair, you're completely right.

And of course, they can't see that.

They see themselves as good guys.

Musk, you know, sees himself as saving the world, but you're completely right.

Objectively looking, why should we trust them?

And as you also point out, their economic interests in this are unbelievable.

I mean, how are they supposed to see this clearly when they are literally making their companies invest tens of billions of dollars into this technology?

And they stand to make themselves many multiples of billions of dollars off the back of this.

So

it wouldn't, in normal world, make any sense at all to have a company say to you, government, get out of the way.

Who the hell does government think it is?

Government doesn't really understand what we're doing.

Let's go.

I think there's another question about whether they're actually right, that it would have been impossible to get an agreement with China.

Now, I can see why they think that, right?

China lied to Obama about islands and the sea.

China often isn't straight with people.

But I believe that if

you had had a serious conversation with Xi Jinping, the number one thing he cares about is the long-term viability of the Communist Party in China.

If he felt that artificial intelligence was a threat to the stability of the Chinese regime and indeed to humanity, he's not part of a death cult, and he also believes that China may be a little bit behind the US in this, then I imagine you could set up an independent verification process.

Why?

Because this stuff can't be done in your backyard.

It requires tens of billions of dollars, enormous compute.

There are basically only five companies in the world and the Chinese government that can do it.

These data centers can be seen from space.

They consume so much energy.

So I think actually if you had moved a year ago to try to be serious about this, it is possible that you could have got these processes in place.

But now, and this is the gloomy bit, we're in a world of Trump, who has just got rid of what remained of Biden's AI Safety Institute.

The UK is a leader in this.

The UK AI Safety Institute is probably the most practical, flexible, thoughtful thing.

Shout out to Rishi Sunak, who actually brought that together.

Shout out to Peter Kahl, who continues to support Navest and that.

But it's increasingly marginalized.

Nobody's paying attention to this really good model that we've got on the UK.

And final question for you.

If possessing one of these things is basically the only way you can have any strategic autonomy in the world,

should the European Union not now be putting everything with the UK and other allies, maybe South Korea and Japan, into developing its own large language model?

Because if you don't have your own large language model, you are completely beholden to the US or China.

And the only thing holding you back now, as the Chinese have proved, isn't actually the scientists and the research.

It's basically the money.

It will cost a lot, but given we're going to spend 5% on GDP on defense, why don't we put half that defense spending into these large language models and have our own independent large language models?

Listen, a couple of things before I come on to that.

And I also want to press you a bit on why you've changed your mind because I can remember when we first started doing this podcast, you were like sort of, you know, really excited about this, really enthusiastic.

But on the energy point, you said that there are these data centers that you can see from space.

Meta are building one that is the size of Manhattan.

We've got 500 already in the UK, and

the indirect carbon emissions of the big tech companies have gone up in the last three years by 150%.

Do you know that by 2030, it's reckoned that these data centers are going to consume the same amount of electricity as the entirety of Japan today.

Ireland,

which has got 80 of these data centers, a fifth of Ireland's electricity supply now goes on feeding them.

So these are, and of course, you don't hear anything from the Zuckerbergs of this world about how they're going to sort of bring solutions to that.

They're expecting everybody else to do that.

So it won't surprise you, note, my answer to that question about Europe is yes.

In the series we just did on J.D.

Vance, when we were talking about that at the end of it, and I was pointing to an interview that Versal von der Leyen, the president of the European Commission, did recently,

and it was fantastic, it was music to my ears.

She said, Europe is still a peace project, Europe is not governed by tech bros and oligarchs, Europe is a place where we don't invade our neighbours and we don't punish our neighbours, Europe is a place of rising life expectancy.

On and on she went.

And if only we could bring that sort of confidence and that sort of vim to, as you say, what is going to be the biggest economic, societal challenge coming down the track, and at the moment, it feels like, and I'm not criticizing Peter Carr, I think he's doing a great job, but at the moment, it feels like the smaller, the middling countries are sort of really fighting to stay involved in a battle that is being fought between America and China, but actually is being won by the tech companies, including in America and China.

So, the dream, I think, would be for the European Union and Britain, which together make 15% of global GDP, right?

It's not that we don't have wealth

and we've just made this big commitment on defense to build their own models and then show that these models from the beginning are engineered to be safe, to be not dominating, and to make sure that the economic benefits of these models are shared with society as a whole.

Create that global standard and demonstrate that building models that are more honest will also be models that are more effective.

In other words, challenge the idea that the only way to get a powerful model is to get rid of regulation and to take risk.

Instead, demonstrate that actually building them from the beginning to be more rigorous and honest will actually make for a better model.

So it's not that you're slowing yourself down.

You're getting something better at the end of it.

Why did I change my mind?

It's been very painful.

I mean, as you know, I'm entranced by this technology.

I use this stuff a couple of hours a day.

I'm very close friends with many of the people that are right at the heart of developing these models, some of whom are understandably very angry with me and saying I've suddenly become a Luddite.

You know, I'm like somebody from the early 19th century destroying industrial machinery because I'm afraid of it.

And I'm having to find a way to say, listen, I completely believe in the benefits of AI.

It is extraordinary.

And I totally believe that the medical revolution that it will bring will transform our life expectancy.

I completely believe that it will generate incredible GDP growth at a time when we need it to wipe out deficits.

I totally believe in its productivity benefits, but the risks are immense.

And

those productivity benefits come sadly with machines replacing humans.

You know, Dario from Anthropic, I think, has talked publicly about 20% job losses.

That is enormous, 20% unemployment.

happening very quickly in economies.

And that's before you get on to these questions of risks and risks that we can't, and that this is where they would argue back against me, right?

So let's say Yoshua Bengio says there's a 1% chance this could kill us all.

There's a 10% chance this could destroy democracy.

Their response is often to say, well, what is that 1%?

How do you know it's 1%?

Why isn't it 0.1%?

Why isn't it 0.01%?

Well,

the answer is we don't know, right?

I think the other thing they'll say, though, Rory, is, yeah, but there's a 30% chance that

we'll eliminate all the sort of terrible diseases rampaging across the poorest parts of the world so that the truth is as you say we don't know um what we do know is that

it is happening it's happening at a pace and a scale that they didn't foresee i mean uh sam altman and we were in davos um and i went to a thing that he was at and he was talking about this thing about power he said it's going to consume way more power than we thought well given there's the climate crisis going on at the same time,

another thing I read said that by,

I think it's in three years' time, two years' time, these data centers are going to require 1.7 trillion gallons of water.

Okay, where's that all coming from?

So

I think these guys are, they feel like they're the masters of the universe.

And I think the reason I'm more worried, and I think this may have had an effect upon you, not me, but the point I'm going to make, is that Musk, I'm afraid, in his flirtation with politics, has given all of these guys a very bad name.

And they're going to have to work a lot harder to get that name back because otherwise, I think people who would normally be sympathetic, like you, are going to start to ask far deeper questions about what they're up to.

Final practical thing for me.

People say, what should we do?

I think let's look at the model of the UK AI Safety Institute.

And it would be good to see Keir Starmer and Peter Kahl lean into this.

So the way to regulate is to define clearly to these companies what we're worried about.

For example, we're worried about the creation of new poisons, bioweapons, set clear metrics for what we want them to be looking at, set clear auditing procedures and transparency.

There's actually a new law coming out in California trying to do this.

get out of, I suppose, where we are at the moment, which is the sense that anyone who's scared is a Luddite and anyone in favor is on the side of progress and find a common ground where we say of course government has a right to be worried about safety and this is a good way of regulating that allows these companies to innovate while also making it clear what it is we're worried about and how we're going to measure and control those things.

Yeah.

My final point before we go to a break and you mentioned the use of AI in Ukraine.

We'll put in the newsletter this really long account of the use of AI in Israel,

And it really

makes the case that one of the reasons there was so many women and children being killed in the early days is because a lot of the killing was being done

with using AI systems to identify where people on their target list were.

So they were kind of using mobile phones, tracking them,

waiting till the person was, you know, in the home, and then bang, there they go.

So this stuff is happening now.

Given the pace at which it's developing, it's going to get faster and faster and faster.

And I've got to say, when that guy, Jeffrey Hinton, first started raising the alarm, I kind of was still in the mode of thinking, oh, well, let's give it time, let's wait and see.

But right now,

having spent a couple of days looking into it since you'd raised the alarm again,

I was pretty alarmed by a lot of the things I read.

Coming from these people who really do understand it.

I don't understand how it works, technically.

They do.

So on that, maybe

let's take a a break.

We'll come back with far better news, Rory.

Younger people in the UK are going to get the vote.

Manifesto promise made, manifesto, promise delivered.

Good for the country, good for the world.

See you then.

See you then.

This is an ad from BetterHelp.

All of us, I think, often do find guidance in very unlikely places, an accidental meeting, seeing someone that we haven't seen for a long time.

But there's something quite different that you can get from a therapist, from a trained professional who has actually had the discipline and the experience to engage.

And BetterHelp has been doing this, finding the right matches between people and their therapists for over 10 years.

They start with a short questionnaire, which helps you connect to an experienced therapist whose expertise aligns with your needs.

And if the match isn't right, you can change it anytime at no extra cost.

It's entirely online.

You can pause your subscription whenever whenever you wish.

BetterHelp has already supported more than 5 million people worldwide, and for many of them, a right match can guide them through not just the moment, but can make a real, real difference.

Our listeners get 10% off their first month at betterhelp.com/slash restpolitics.

That's betterhelp.com/slash restpolitics.

Welcome back to The Rest is Politics with me, Rory Stewart.

And with me, Alistair Campbell.

And Rory, before we get on to votes for 16 and 17-year-olds, just a reminder to people: Gary Stevenson, we interviewed him over two parts, two episodes.

The first one dropped on Monday.

But members, Trip Plus members, can get access to part two anytime they want.

Everybody else, it'll be out next Monday.

And also, I have had a little chat these episode six of the JD Vance mini-series with David Frum.

Great guy.

We've interviewed him before.

He was effectively Vance's boss for a while.

And talk to him about what he thought about the series, more importantly, what he thinks of Vance and what he's going to do in the future.

And that's going to be out on Friday.

Very good.

So if you'd like to join the Restless Politics Plus, we'd love to have you.

You can sign up at therestlistpolitics.com and get, for example, early access to episodes and the special member series, live events and many other benefits.

Now, Alistair, we are now getting on to votes for 16 year olds.

And this is one where I think we have had an agreeable disagreement.

So let's play this out for people.

So you've been an absolute unapologetic champion of votes for 16 year olds.

I've been a little bit more skeptical.

I used to go around Cumbrian schools, and as you can imagine, every Cumbrian school, everybody really wanted votes for 16, 17-year-olds.

And I went went around alienating my future voters by telling them I didn't really support it.

I guess in the end, it's a question of balance, isn't it?

How young you want to go.

I suppose the vote used to be 21 and then it came down to 18.

And it's about balancing

the sense that in many ways a 16-year-old is a adult.

17-year-olds can go off and join the army and stuff, pay taxes.

In other ways, other bits of society don't fully consider them adults.

So there are restrictions, for example, on access to alcohol, cigarettes.

There's evidence about the prefrontal cortex, particularly of young men.

Sarah Jane Blakemore, for example, has done research on this.

How did you weigh this all up in your head?

And why did you decide that you wanted to get to 16 rather than 14 or 17?

Well, I'm not convinced that I wouldn't actually go lower.

But yeah, but

I think I told you last year, I went back to my kids' primary school recently and this kid came out to me and he said, you talked about me on your podcast, didn't you?

And I said, did I?

I said, yeah, I'm the one last year.

You were talking about lowering the voting age.

And I said, why can't you lower it to two?

And you said that on your podcast.

So there's a primary school kid listening to our podcast, Rory.

No, I think there's six things about right.

Why did I fall upon that?

Why did I think this was a good thing to do?

One, because I think you agree with me.

Our politics is in a real state.

our political debate is in a real state.

I probably have more appreciation of the value of young people than a lot of older people do.

I like going to schools, I like talking to kids, I find passion in there, I find a desire to hunger to learn.

Yes, you find people who are cynical and you find people who don't care and all that, but you find them in every generation.

I have a feeling with our country, this is, I'm going to get a lot of our older listeners will not like this, but I actually think we get less curious and less and more cynical the older we get.

I try to buck that trend.

I hope I'm not more cynical, but I just think that the

other thing, Rory, we've had so much negativity around this government.

This to me is a positive thing to do.

I don't think the Tories will reverse it, and therefore I think it's a kind of genuine historic constitutional change.

And the other, the big clincher for me was the Scottish referendum, where I really, really feel that the sixteen and seventeen year olds added something major to the debate.

And I hope that happens now.

And by the way, lots of the people, when I've said I'm in favour of this and I did some interviews the other day and people say, oh yeah, because you think all the young people are stupid enough to vote Labour.

I actually think this is a huge risk for Labour.

That's why they're to be commended even more for doing it in my view because I think that

they're not very popular with young people at the moment over Gaza, over tuition fees, over housing, lots of different issues.

I think there's a real danger with the sort of polarisation.

We've talked before about Farage's dominance of TikTok, for example.

That actually is a danger that, you know, this disproportionately favours parties on the extreme.

But that's why I would say, if I was the chief whip of the Labour Party today and the campaign manager, every Labour MP and candidate between now and the election do a lot of your campaigning in schools, not least because it'll give you more energy for the campaigns you do elsewhere.

One question to you.

Do you feel personally, for example, example,

with Grace, that

when she was 16, there was any real difference to the kind of political judgments she would have made compared to when she was 18?

Did you feel that she developed a richer understanding of politics?

What I love about voters and the reason I believe in democracy is I always felt in Cumbria that people, regardless of their background, had a really deep understanding of their own reality.

They had a really practical understanding of their own place, their own job.

And actually, one of the reasons why democracy works is that when people are voting, if they can connect it to their everyday experience, there is incredible wisdom.

They come up with the right answer.

The issue with a younger person maybe is that they may be living with their parents.

They may not actually have a job yet.

They may not be paying taxes.

And therefore, when they're making decisions on government spending, planning, et cetera, it may not be kind of rooted in reality in the way that my kind of ideal Cumbrian farmer kind of feels every minute of their day, the consequences of these things.

Yeah, the only thing I say to that is people, if they're still in school,

16, 17, 18, they're spending more of their time being curious.

They're spending more of their time kind of thinking through problems.

Look, let's be honest, Roy, you've been an MP.

You've done lots of campaigning.

You meet all sorts of people.

And let's just be frank about this.

You regularly meet people and think, oh my God, did they really just say that to me?

Oh, my God, did they really think this?

Oh, my God, do they really think it's a fact that etc etc etc you get that of all generations i'll tell you the other thing i think our politics we're going to talk in question time about the election in japan one of the reasons the ruling party in japan is doing so badly is because young people are turning to the right-wing populist parties because they think that the mainstream parties are too focused on old people

and not enough on young people.

And I think this is a sign of saying, we're going to take your voice seriously.

We're going to listen to you.

I really hope this changes the way that we debate politics.

I really hope that it changes the way that MPs campaign.

I really hope that it changes the way that the media think about politics.

Don't patronise them.

They're now on a par with us.

They will get the vote, same as you and me.

And the point about Grace, Grace, I remember, used to be against this.

She used to say, oh, God, dad, my friends don't know anything about politics.

But I don't care.

This is an opportunity for us to regenerate political debate.

And I would put it, as you you know, alongside compulsory voting and compulsory political education in schools.

But I know that that's for another day.

Well, let me finish with that then.

Take your inspiration.

I would be most excited if I felt that this was the beginning of real constitutional change.

I think the big challenge for our generation is to rebuild trust in human institutions.

Big emphasis on the human, given all this conversation we've had around AI.

How do we trust government again?

How do we trust our democracies again and believe in them?

So I would really think that this is the moment to be serious about what's going wrong.

How do we get governments that deliver and are effective?

So how do we get this balance between the process and regulations we want to protect us along with governments being able to drive forward?

Secondly, how do we think about all the things that are wrong with the incentives of the political parties, all the stuff that I saw in parliament, which really disgusted me about the way that MPs thought, the lack of seriousness, right?

So that's one of the reasons I believe strongly in citizens' assemblies, because citizens' assemblies are

like juries.

They're randomly selected from citizens, and by definition, they don't have party political interests.

They're thinking about these things

without thinking all the time, how do I stick it to the opposition?

What does this opinion poll say has to tell me?

Local democracy, how do we really drive power down to the localities, give financial power?

This isn't just about, you know, a mayor of Cumbria, but a mayor of Penrith, the sort of French model, who write down at the local level.

And how do we tap the fact that we have people in our countries who, on the one hand, have never been kind of so educated, so well-traveled, so informed about the world, but on the other hand, have never been so despairing about the future, angry with their governments.

And if we're going to address populism, if we're going to address these huge challenges, of which I now think AI is one of the biggest challenges we're going to face globally, we've got to rebuild the quality of our system.

And to rebuild the trust in our institutions, we have to rebuild the way those institutions work, which I think goes a long way beyond 16-year-olds.

I totally agree.

We both said this ahead of the election.

I really think there's a space for a huge review of the way our politics and our constitution works.

But what I like about this, and you know, there's so many people when it it happens in who the hell, nobody ever told me about this.

This was in the Labour Party manifesto.

This was something they said they were going to do.

So let's not get too aerated about that.

And the other two points to make, just on the polling, those who vote young tend to carry on voting.

So start them young, I say.

And the other point to make is that the support for extending the franchise to 16 and 17 year olds in Scotland among adults grew

after that referendum and after subsequent elections.

So I think they've seen that it's a positive within the body politic.

Finally, the electoral system, where I think you and I also agree.

I'm increasingly in favour of a New Zealand-style electoral system where you balance proportional and first-past-to-post.

We're now in a world where a party that gets just over 30% of the vote can end up with a big majority in parliament.

Now, you don't worry too much about that when it's Keir Starmer, who's a relatively middle-of-the-road old-fashioned politician, whatever you think about him.

But we're now getting into a world where Nigel Farage is in many, many opinion polls, certainly well ahead of the Conservative Party and probably ahead of Labour, a world in which it's quite possible that he could get 32% of the vote.

And then you'd end up with him dominating parliament with 70% of people not having voted for him.

So I'm increasingly convinced that as social media, as our economies, as all these things drive us to more and more extreme solutions, there's more and more of an argument for an electoral system which tries to deal with the danger of fringe parties exploiting a first-past-the-post system to take majorities.

Well, you're supported in that view, of course, by Nigel Farage, who used to be massively in favour of proportional

representation and now is no longer.

I said last week, one tactic I think, or one approach Labour should take is actually constantly to call out Farage on his one achievement, which is Brexit.

But there was some amazing stuff this week on Farage.

I mean, he honestly, he's, I think he's, yes, he's a good communicator, but I think he's a great target for another party.

And the Tories have got to start attacking him properly, and so have Labour.

I mean, he came out with stuff on crime yesterday, and okay, he gets his front pages in the mail and the telegraph about he's building new prisons left, right, and centre.

30,000 new police officers are going to come on board.

He's going to halve crime in God knows when.

It's classic Trumpian populist bullshit, right?

But the argument has got to be won over time.

And I think that he's now, I get a sense that Farage is becoming a little bit more vulnerable.

The media is starting finally to put a little bit more.

He's still got his fanzines like the Mail and The Telegraph, but the media is starting to put him under a little bit more pressure.

But ultimately, this now is a Labour doing this should be a challenge to them.

How do they now start to win the battle for young people on social media?

That means improving your social media.

It means doing it better, doing it more often, being inside the feeds of young people much more than they are because Farage is all over them.

But secondly, it means having this national story about a national mission and showing that you can deliver it.

And that, frankly, is as important to young people as it is to older people.

Shout out for my final comment to Ansha Rainer, who's trying to bring back in local mayoral elections the

transferable vote vote system that used to exist in the London Mayor, where if you don't get 50% of the vote in the first round, there's the second round.

That's what allowed Macron to come through in the French presidential system.

And I think it was a real tragedy that we got rid of that.

And also a sense that this is something I've changed my mind on.

I was passionately pro

first past the post.

I believed as a constituency MP that the constituency relationship was so precious.

And I felt speaking to Spanish colleagues who were in a full full proportional system that they had very little relationship with their constituents because they were just on a list and it made parties too powerful.

So I think it's really been the New Zealand experience that you can make a balanced system and get a lot of the benefits of both systems that have driven me to change others.

So Rory, in the last two weeks, you've changed your mind on defense spending, AI, and first past the post.

I think we should ask our listeners the final thing, write in and tell us what you think is the next thing Rory's going to change his mind on.

The flip-flopper in sheath.

Very good.

Okay.

And let's see if he's changed his mind on any of the things we talk about tomorrow.

We're going to talk about Israel, Syria.

We're going to talk about young men and authoritarianism.

We're also going to hear that Rory has seen Superman, which I haven't.

As we said in the main podcast, we're going to talk about Japan.

So lots and lots and lots to talk about.

We'll see you tomorrow.

See you tomorrow.

Bye-bye.

Bye.