Hosted by Christina Ruffini and Sir Richard Dearlove
Guest
Anne Neuberger
Anne Neuberger is a former Deputy National Security Advisor for Cyber and Emerging Technology under the Biden administration. Anne Neuberger explores the growing role of artificial intelligence in defense systems and why China is the leading cybersecurity adversary.
Episode Summary
What is it like to meet with Russian leaders in Moscow? Will President Trump's phone call with President Vladimir Putin help end the war? And why is AI now the hottest tool in cybersecurity?
In this week’s episode, Christina Ruffini catches up with co-host and former head of Britain’s MI6, Sir Richard Dearlove, to unpack the global implications of President Donald Trump’s phone call with Russian President Vladimir Putin earlier this week. Sir Richard shares exclusive insight into his meetings in Moscow and what Ukrainians are telling him about ending the war. He also reveals what he believes is Putin’s true vulnerability. Later in the episode, Sir Richard sits down with Anne Neuberger, former Deputy National Security for Cyber and Emerging Technology under the Biden administration. They explore the growing role of artificial intelligence in defense systems and why China is the leading cybersecurity adversary.
Episode produced by Situation Room Studios. Original music composed and produced by Leo Sidran.
Listen & Subscribe
- Watch on YouTube
- Listen on Apple Podcasts
- Listen on Spotify
Transcript
COLD OPEN
Trump: I'll be speaking to President Putin on Tuesday. We wanna see if we can bring that war to an end. Maybe we can't. Maybe we can't, but I think we have a very good chance.
INTRODUCTION
Raffini: Welcome to One Decision. I'm your host, Christina Raffini. This week, we're joined by co-host, Sir Richard Dearlove, of course, former head of Britain's MI6. And there's, as usual, as there seems to be more and more, a lot to unpack in world events as everyone currently waits to see what exactly was discussed in that notorious two, maybe three hour phone call between US president Donald Trump and Russian president Vladimir Putin earlier this week. Here's what we think we know. Putin allegedly — allegedly — a ceasefire, agreeing instead to halt attacks on Ukrainian energy facilities but only for thirty days. Just hours after the phone call a Russian drone hit Ukrainian hospital. So of course the bigger question is what does this kind of limited ceasefire mean for the ongoing conflict and how does this position Ukrainian president Volodymyr Zelenskyy in any future possible negotiations? Alright. So Richard is here. He just got back after a very long drive, but he's logged on, and we're gonna get into it. First of all, what is your take in general on a three hour phone call with Vladimir Putin, and how do you think that goes?
ANALYSIS WITH SIR RICHARD DEARLOVE
Dearlove: Well, they clearly have an understanding which allows them to talk to each other. And given, you know, Putin's isolation internationally, it must warm his heart to be talking one to one to the president of The United States. But on this occasion, it's extraordinary that this has actually taken place. I mean, none of us could have guessed it. None of us could have predicted it, but it's happening. Now sometimes, to talk to your enemies is a very good idea, but, as it were, the existential question is what is the nature of the conversation? Because, I mean, I've been following the media and listening to various people being interviewed in Moscow. The distance between the Ukrainian position and what we believe the Ukrainian position is likely to be if it evolves at all and the Russian position is infinite, absolutely infinite. And here's Putin saying, you know, no foreign support, no intelligence support for Ukrainian military, and that's the whole basis of the European support for Ukraine. So how on earth do you bridge this gap? And, you know, Trump tells us he had a very good conversation with Zelensky, that this is a very constructive conversation with Putin. Of course, it was. That's how it's gonna be portrayed. But have they actually made any significant progress? And I think we would all like to see an end to a very bloody, costly war. But, you know, this isn't a real estate deal. It's trying to end a major European war. And I have been talking to a few Ukrainians very recently, and what they're saying to me is, even if it were a bad deal, they're gonna fight on on their own. The determination of the Ukrainian people is extreme. I mean, they've paid such a high price already. They're not gonna stop. Even in the most adverse circumstances, they're not gonna start sort of surrendering. And, I mean, Putin is still virtually, as far as I can see, requiring more or less a complete, you know, surrender on the part of the Ukrainians to everything he wants.
Raffini: So how is Trump gonna square that circle? It seems like even very publicly, the Kremlin's position is basically what Putin has been wanting all along. Right? No intel sharing. No NATO troops in Ukraine. Well, that's how we got here in the first place. It doesn't seem like he's moved at all or is inclined to. You mentioned that it's not a real estate deal. Tom Friedman in the New York Times mentioned that, you know, it took secretary of state Henry Kissinger over a month of shuttling back and forth between Israel and Egypt to, you know, stop that war in 1973. So he says Trump couldn't sell a hotel that quickly unless he was giving it away.
Dearlove: No. It's crazy. And I absolutely agree with you. And let's face it. I mean, Kissinger was a master of the art of diplomacy and fixing things. But if you read any of Henry's books, I mean, hours, months of preparation for these very complex initiatives and breakthroughs, like, you know, speaking to China as he did and developing a new relationship with the Chinese. And the idea that Trump can sort of drop out of the sky and just suddenly fix it. No. And, I mean, I think he's now beginning to encounter the reality. Look. He can bully Ukraine, which he certainly did, into saying, we'll accept a thirty day ceasefire. No conditions. But Russia? It's an entirely different question. So, I mean, I would love to see a transcript to the conversation.
Raffini: Yeah. I don't think you're getting one of those. You don't think so?
Dearlove: Well, I just wonder what line they did it on. They must have done it on some — it's interesting to imagine, you know, how secure was the line between the two of them? Did any of those sophisticated intercept services in the world intercept the conversation? Probably didn't. But, I mean, if I were sitting in Whitehall supportive of the Ukrainians, I'd love to know exactly what happened in the conversation to try to get the dynamics and understand the chemistry of this peculiar situation. You may not have heard it, but there was an interview on the BBC with this dreadful man, Markov. You know, what he was saying was so extreme, you know, Nazi fascist regime, you know, this isn't an independent country. And, I mean, he does more or less represent the Kremlin view. And then you've got Peskov also, the spokesman, saying, you know, the president of The United States and Putin, they trust each other. I mean, where are we heading with this? I'm hoping that the sensible voices in the Trump administration — and let's be fair, there are some very sensible voices in there — that they are heading in the constructive direction. But at the moment, I'm really quite worried.
Raffini: Well, talk to me a little bit about what it's like to deal with the Russians.
Dearlove: I was head of service during a period when there was a brief honeymoon post 9/11 before, you know, things went belly up when Blair and the rest of us thought, well, here's Putin. He's a new leader. Maybe we can have a different sort of relationship with Russia. And we did try quite hard, and we went to Moscow quite a bit, and we had some interesting meetings. But the trouble is that you feel a lot of the time that you're being set up for a fall. The best advice I ever got about Russia is, once you're in Russia, nothing happens to you by accident. So the more bizarre the event, you know, it didn't happen by accident. And, you know, we were made to wait for meetings. But when things got icy, which they did in my final visits to Moscow, I mean, it was just bizarre. Things went wrong. People didn't turn up. And, I mean, it was all deliberate to make you feel suddenly that you were unwelcome.
Raffini: We were talking a little bit about the relationship between Trump and Putin and this level of trust that Trump seems to have in Putin, unclear how Putin feels about Trump. We spoke to Fiona Hill in a previous episode. She's a former senior director for European and Russian affairs on the National Security Council, and she also worked on the Trump administration. She had some interesting things to say about the Putin-Trump relationship on the podcast. Let's take a listen.
Hill: Putin is pretty confident, given his experiences with Trump in the past, that Trump will be quick to try to resolve the whole conflict and, you know, war in Ukraine in his favor. You know, obviously, he has had Trump's number for some time. He knows how to manipulate him. You know, he was being very good at the art of flattery with Trump. He sees Trump as an asset in many respects because, although Trump can be extraordinarily unpredictable and, you know, Putin probably has to tread very carefully in fact not to insult him and not to kind of cross lines that Trump, you know, lays for himself, Putin is, you know, pretty confident that he can stoke up the culture wars here, there, and everywhere, and just with a little bit of deft use of political influence operations and propaganda he can, you know, keep things sort of moving in his direction. It already is moving in his direction in any case.
Dearlove: I think there's a certain naivety and, let's say, intelligent simplicity about Trump. You know, I'm still hopeful that somewhere in this extraordinary series of events, there might be, you know, a positive outcome because there are super intelligent people in the Trump administration who, okay, are themselves worried but feel that there is a strategy and a direction. But at the moment, I can't see what it is, and I'm worried. But I'm prepared to remain, to an extent, open minded before this finally plays out. There are things that have to happen, because I think the only way that Putin comes to negotiate seriously on a ceasefire is if things start to turn adverse for him. And at the moment, there's absolutely no sign of that at all. Putin's vulnerability is actually the Russian economy. So there's no lever you can pull that has an immediate effect, but there are plenty of levers you can pull which, in several months, begin to seriously affect Russia's position. And let's face it. They're not in a happy place in regard to this war. They're in a mess. I mean, we tend to think, well, you know, Russia holds all the cards. It's not as simple as that. They're in a pretty bad place as well themselves. So the Russians do have an incentive to reach a deal, but at the moment, they're only proposing a deal which gives them everything they originally asked for. And I don't think that's a viable position. Let's see what happens next.
Raffini: All right. Look, speaking of looking to the future, we want to turn now to a conversation that you had on your own with Deputy National Security Advisor Anne Neuberger. And as I understand it, you talked a lot about artificial intelligence and cybersecurity and how that's fitting into future combat roles and future intelligence gathering.
Dearlove: That was a tough conversation because she is one bright, well informed lady who has done, you know, a really key job. Okay. She's moved on now from the NSA. But I mean, it's clear that AI is going to transform aspects of, let's say, technical intelligence gathering. But I think I understand the basics reasonably well. And I think some of the stuff she said was extraordinary, and, yes, it's gonna play a huge role. And it will make it much, much easier to capture and interpret knowledge from these massive data flows, which are so huge, you can't do it in any other way than through using AI. And I mean, I think the key question I asked her in the interview was the relationship between AI and AI. And what I mean by that is, can you use AI to control AI?
Raffini: Well, let's take a listen right now. Here's that interview.
INTERVIEW WITH ANNE NEUBERGER
Dearlove: And welcome to One Decision Podcast. I think my first question is, where at the moment do you judge cybersecurity to be? I mean, how well are we doing? My own experience is, I think that cyber sophistication has run ahead so fast that we're always patching to catch up with this sort of exploitation of the capability, and that the security so often is applied in a sort of a secondary fashion. I would hope now that the world is changing and people understand the importance of putting cybersecurity first, but what would be your observation on that point?
Neuberger: So Richard, it's great to be here with you. I think you captured it very well. A lot changed in the last four years, notably in The United States, where The US began using a more powerful set of tools to guide national defense. The first part — which, and I'm happy to describe this more, but I'll briefly note. The first part was actually requiring regulations that required minimum cybersecurity for the most critical networks that Americans rely on every day, from ports to airports. The second was using the power of government procurement. The government spends $51,000,000,000 on tech every year to say we're going to apply that to require cybersecurity so that we are putting our money where our mouth is in incentivizing more secure tech. And finally launching innovative public private partnerships like the Cyber Trust Mark to secure consumer connected devices like baby monitors and home alarm systems in a creative way. So those three are examples of the just more aggressive approach in The United States and in many countries around the world. But I lift up The United States efforts over the last few years to counter the dynamics on the attack side.
Dearlove: But I mean, this is very much a layman's question from my side. Can you sort of apply this layering of AI on a continuous and contemporary basis so that you are immediately alerted to when, you know, you've got a hole in your defense or when something odd is happening. I mean, my question is, is it contemporaneous or is it an analysis after the event?
Neuberger: Richard, what you outlined is the absolute goal. It's where we need to be. You want it contemporaneously because, you know, the most important thing is to actually find an attacker before they found what they want on the network and taken it out, whether it's corporate secrets or whether it's private information. So that is exactly where we need to be. I think the tech's a mixed bag still. It's not quite there yet in terms of actually being able to do so. But a number of companies are looking at it. One of the last steps, again, in the Biden administration was releasing an executive order that put that as a priority in terms of research and development work to get that there. I would note one area where we see a lot of promise is in code and in generating code. So while there's a lot of AI efforts being — and good products that can generate code — what we'd love to see is that generate more secure code, actually train code generation models to look at commonly known vulnerabilities, commonly known attack techniques, and correct them in the code that they're generating. So that's a call to action for using AI in a way that would be tremendously valuable in new code being generated.
Dearlove: Yeah, that's fascinating, because I've been talking recently, just coincidentally, because I know quite a lot of clinicians, and it seems that AI is absolutely brilliant at diagnosis. In fact, in many cases, far better than any other methodology. Even the best professors get nonplussed by the speed of AI's diagnosis. So, I mean, it seems to me you're suggesting the same level of sophistication when it comes to the cybersecurity and the ability to analyze very fast when something is amiss. Taking that sophistication as the level of our discussion, where is the most serious threat now coming from?
Neuberger: The most serious threat now is coming from China, in that China views its active, well resourced, sophisticated cyber program as a tool of its geopolitical goals. So we see China targeting, you know, countries in its region, as well as countries in Europe, The United States, in a number of ways. First, still stealing intellectual property, but second, not only hacking governments for routine espionage, but more importantly, prepositioning in parts of countries' critical infrastructure, water systems, ports, power systems, we believe — we are concerned — with the goal of disrupting or degrading the operations of those critical services in a period of crisis or conflict to put pressure on a government. So that, you know, people in the country say to the government, why are we getting involved in a crisis halfway around the world? You know, let's focus at home. So that's our most significant concern about China. I will note, as a quick corollary, one country that also concerns me a great deal is the degree to which the North Koreans are using hacking to get around sanctions. Really, the most spectacular, in terms of size, hack of a cryptocurrency entity, the North Koreans just recently conducted, that stole $1,500,000,000. So that's another key. China is kind of prepositioning for its future goals. North Korea is actively using hacking in cyberspace to finance its dangerous missile program in a way that concerns us.
Dearlove: But I think what worries me in terms of what you're saying is that if you take an autocracy like China, presumably it can use its infinite access to databases, because it has no privacy models at all which are applicable to the state, which must give it enormous potential power in the medium to long term in building its own AI models as a means of offense. I mean, do you agree with that, that autocracies have a built in advantage in that respect?
Neuberger: I do. Because, you know, our civil liberties and privacy, which are a key part of our life and something we deeply, deeply value, also constrain us in the context of AI. For example, in building computer vision models of facial recognition of cameras. You know, we want to be protected from that. We want to require our government to ask our consent first. You know, a country like China, which deploys technology first at scale — you know, you visit China, wherever you look, there are cameras. They also, as an authoritarian government, are not particularly concerned about their citizens or frankly tourists or visitors, civil liberties and privacy, and can use all of those images to train advanced computer vision facial recognition models, whether that's for surveillance for criminal purposes or for their national security spying or espionage purposes. But I think, you know, that should be a call to action for us to not lag because we say, you know, how do we — there are difficult new policy issues. They are. There are difficult new policy issues. We have to figure out the most important national security goals that we need as a country, as allies, where AI can make a big difference. And then ensure that we talk transparently with our publics about what AI can do, why it's needed, and explain clearly what we're doing in the context of that. So that's a really important thing. So we need to be first, you know, first to really use AI in line with our values. We cannot leave those values behind given the power of AI and given the import in the national security arena.
Dearlove: I'm interested also, I think all our listeners would be interested, in the extent to which the big tech companies are sort of influencing the issue we're discussing, whether it's Musk or whether it's Zuckerberg. I mean, these companies now are so dominant and important in the cyber area generally, or let's say, the information security. How important are these companies in, as it were, driving the issue of security quite apart from what's happening in government?
Neuberger: They play an absolutely important role because these are their platforms, and they have the best sense of the code and how to ensure that it's secure so users are secure from hacking while using those platforms. You know, so that's one of the reasons why throughout my career, I've put a great deal of work into building just creative ways to bring the intelligence community, notably NSA, and key technology companies together so that the intelligence community was sharing focused threats, so companies have a better sense of what they need to defend against. And technologists across the public and private sectors can put their heads together on how to defend against those threats. So that ranged from bringing together a whole group of the AI LLM companies. We brought them together at the White House with the National Security Agency, CISA and others to provide a classified threat brief, to stand up a Slack channel where companies and government can share information regarding threats they see, and most importantly, the best way to defend against that. That extends all the way to inviting the telecom companies in to discuss Chinese hacking of telecoms in the fall of 2024 that had been going on for a number of months. Again, to bring government and private sector experts together to say how does one defend against these threats in the most effective way? Because clearly, AI companies, telecoms are going to remain targets because their platforms are being used by so many companies, by so many governments. And we wanna bring the best talent to bear in government on these problems while noting that the platforms are owned and operated by the private sector, and they have the final ability and capability to actually be keeping them secure.
Dearlove: It isn't a great secret because it got widely publicized, but I had my Proton email hacked by the Russians, and a lot of the content of my private email spread around the Internet after the event. And there was a political element to it because I was very heavily involved in the Brexit debate. I mean, have you any advice for people listening? Because, I mean, I said it was a bit like sending the pub soccer team out to play Manchester United if you were an individual and you were up against the Russian state.
Neuberger: I love that. It's a very apt comparison, and it's very well said. And it's one that everybody can say, oh my. There are a set of practices. They're sometimes not easy, but a set of practices that does make one a harder target. Things like multi factor authentication. A second, and not just a text, a second hardware factor that is needed in order to log in, for example, to your email. Similarly, keeping systems patched continuously, upgrading them, turning off and on your phone regularly, obviously using the best cyber defense. And frankly, one of the reasons that we've been, from a government perspective, putting so much work into working with companies is to say what's occurred in cyberspace, whereas, as you put it, each individual feels like they're up against a powerful government's advanced capabilities, it's not a fair fight. Companies can do more to build products more securely so that that fight is more evened out. Even though, as I noted in the first thing we talked about, it's far easier to attack than to defend. So I think for individuals, first, using multi factor authentication because passwords are dead. Second, backing up your most important information and having that as an offline backup. So for example, if somebody conducts an attack against you that encrypts your computer and tries to ransom that, have you pay a large sum of money to get that data back, you have an offline accessible copy that's made available to you, particularly for your most sensitive financial, personal, medical health records. Patching devices regularly. And if you're a corporation, there's a few other things. Monitoring networks continuously and using AI and adopting that to find what is a baseline configuration, what's baseline, what's potentially anomalous, exercising an incident response, so that you would know what will you do if you're hacked so that you can more quickly recover.
Dearlove: But I mean, how do you view the problem now looking forward, let's say, for the next couple of years? Where are the biggest vulnerabilities? Where are the biggest opportunities? I mean, does it all circulate around AI? Is that the sort of focal point? Because you've written a lot about how AI is changing the cyber world.
Neuberger: In The United States, over the last few years, there's been a much more aggressive approach in three domains. One, regulating the cybersecurity of critical infrastructure that if disrupted would affect every American's life. It used to be voluntary, and frankly many companies didn't have cybersecurity to where it needed to be because it was voluntary. At the end of the day, it's an expense. It's a cost. And if they're required to do so — and for that I mean again the critical companies — if they're required to do so, they'll take those steps. In The United States, the Biden administration put in place requirements for pipelines, power, water, airports, and initial steps on telecom in those final months for just that reason. It covers roughly 250 companies. And frankly, when you look at pipelines, which was the first sector, today 100% of pipelines meet those standards, and they have not reported a cyber attack since that time. So that shows that one can make it riskier, costlier, harder for attackers. One can ensure that the country's national infrastructure is where it needs to be. And as countries, we're partnering together on that. That's the first step. The second is building creative innovative partnerships to ensure that governments are sharing information with the private sector. I talked about our work in the White House to invite AI companies in. I talked about our work to invite telecom executives in. And certainly, we do that internationally as well. The White House launched the Counter Ransomware Initiative in 2021. It grew to bring 70 countries together annually in person and regularly virtually on a whole set of efforts to secure cyberspace, to help each other, to be available in the case of an incident, and to really work to make it harder to launder some of the crypto proceeds that are used to make ransomware attacks profitable, to do that via crypto. And then the final piece is voluntary creative public private partnerships like the labeling initiative that incentivize companies to build more secure products for the connected devices at home. So that's the first set of work on defense. The second set are more aggressive offensive partnerships to show that there'll be repercussions and consequences. And the third step is using international norms to make clear that certain sets of activity in cyberspace are unacceptable. And as a community of countries, we'll call that out, and there'll be diplomatic consequences as well. Just about a number of months ago, The US went with about 50 other countries to the United Nations to call out ransomware attacks against hospitals, to make clear that that was counter to international norms and that we considered it such a critical issue that we would bring it to the UN Security Council within a broad set of countries to put that on the agenda. So those are the three areas that I see significant progress and continued work needed to continue to make life more difficult for adversaries in cyberspace.
Dearlove: Can AI be built in such a way to defeat the AI threat? Can we use AI to control AI?
Neuberger: I think it takes AI to beat AI. Much as an offensive AI system, for example, in the context of cybersecurity, can be used to find vulnerabilities and to craft a multistage attack that's hard to recover from, AI on defense can first be used to step ahead to also find those same vulnerabilities, to accelerate building out patches, deploying them, and then to continuously red team an important network yourself. We need to be using AI on offense to check our AI on defense. And we need to be using AI on defense to make ourselves harder targets. I think it is one of the most important technologies to even out the significant gap today between defenders and attackers.
Dearlove: Well, can I ask you, you know, whether you had a personal decision on a particular attack or issue which you can talk about from your experience?
Neuberger: After president Biden and president Putin met in the summer of 2021 following a significant ransomware attack by individuals based in Russia that disabled a major pipeline in The United States for almost a week, the two presidents established an experts group to discuss ransomware attacks in cyberspace. And I led The US side. You know, these were virtual meetings. And I recall the first meeting when the screen came up. And on VTC, on secure video conference, was the Kremlin and my counterpart. You know, my first — I must say on a personal level, my father and grandparents came as refugees from Soviet occupied Budapest following the Hungarian Revolution in 1956. But for a moment, as I looked at the secure screen and the Kremlin logo and the white painted wood with gold gilt, it felt very much as if our family's life had come full circle. As my family that had fled Russian communism, the other side of my family that had fled Nazism, was in The US in this free environment. It made me feel very grateful to be on that side of the screen.
Dearlove: Anne, thank you very much indeed. And I very much appreciate our discussion.
Neuberger: Likewise.
Raffini: Alright. That does it for this week's episode of One Decision. I'm Christina Raffini, of course, here with my co-host Richard Dearlove and special thanks once more to Anne Neuberger. If you like this week's episode be sure to check out that previous conversation we mentioned with Fiona Hill. She's the former senior director for European and Russian affairs on the National Security Council and also worked under the Trump administration.
Hill: Everything that Trump surrogates or Trump himself, you know, says about NATO, about Europe and European security, about world and global affairs, about Ukraine, everything that's happening on Capitol Hill. I mean, for Putin, this is just, you know, for him, a sign that, again, everything is gonna rapidly switch in his direction and his favor.
Raffini: As always, for more content and key discussions with global leaders, be sure to check out our website, onedecision.com, and of course, our YouTube. I'm Christina Raffini. We'll catch you next time.
