Financial institution
Crypto business
Home / Resources / TRM Talks / Cyber's AI Moment with Nick Leiserson

Nov 18, 2026 - 39mins

EPISODE 124

Cyber's AI Moment with Nick Leiserson

With Nick Leiserson,  and ,  and  and

Nick Leiserson has spent his career trying to see around corners in cybersecurity — from Capitol Hill, where he helped write the National Cyber Director Act, to the White House, where he worked alongside the nation's first National Cyber Director, Chris Inglis, on critical infrastructure policy. Now President of the Center for Advancing Cybersecurity at the Institute for Security and Technology (IST), Nick joins Ari Redbord, TRM Labs' Global Head of Policy, to talk about a moment he says is unlike anything the field has seen before: AI accelerating both attacks and the exploitation of flaws that have sat quietly for years.

Nick walks through the PHP vulnerability that let an AI model instantly weaponize a bug engineers had known about but shelved as too hard to exploit, and a certificate authority's signing error in India that could let attackers impersonate trusted websites for years to come. He and Ari trace how ransomware has evolved since Colonial Pipeline — from attacks that shut down critical systems to extortion schemes built on stolen data — and why that shift, while still dangerous, counts as progress. Nick also details how blockchain intelligence is increasingly fused with threat intelligence to identify the shared infrastructure ransomware affiliates rely on, giving defenders a path to disrupt many actors at once.

On his wish list for policymakers: Congress should start now on incentives to secure the flood of code AI agents will soon be writing, while the White House and CISA focus on today's urgent software-quality failures. And as "target rich, cyber poor" institutions — schools, hospitals, water utilities — face a widening gap between attackers' capabilities and their own defenses, Nick argues the only way forward is treating cybersecurity as a team sport between government and industry.

Click here to listen to the full episode: Cyber's AI Moment with Nick Leiserson. Follow TRM Talks on Spotify to be the first to know about new episodes.

Ari Redbord (00:01):

I'm Ari Redbord, and this is TRM Talks. I'm Global Head of Policy at TRM Labs. TRM is an AI investigations and threat intelligence company combining on-chain and off-chain data to combat modern crime and build a safer world. We use our AI investigations platform, Ardence, along with Chainabuse and the Beacon Network to bring together the public and private sectors to defeat scam networks, ransomware gangs, terrorist financiers, and other threats.

(00:35):

Prior to joining TRM, I spent 15 years in the US federal government, first as a prosecutor at the Department of Justice and then as a US Treasury Department official, where I worked to safeguard the financial system against terrorist financiers, weapons of mass destruction proliferators, drug kingpins, and other rogue actors.

(00:56):

On TRM Talks, I sit down with business leaders, policymakers, investigators, and friends from across the blockchain, AI, and technology space who are working to build a safer world.

(01:12):

On today's TRM Talks, I sit down with Nick Leiserson of the Institute for Security and Technology.

(01:21):

But first inside the lab where I share data-driven insights from our blockchain intelligence team.

(01:28):

On today's Inside the Lab, leaked documents on Russia's A7 network gave TRM, working with the Open Source Center, a window into an entity its own leadership has publicly denied is even a crypto operation. Ilan Shor, the sanctioned oligarch behind A7, told an audience last year, quote, "We are not a crypto mechanism."

(01:52):

TRM's analysis tells a different story. The leaked documents estimate A7's on-chain volume tops a hundred and sixty-six billion dollars. TRM's own tracing found a hundred and seventy-six point six million in direct exposure connecting A7 to Iran's IRGC, North Korean state actors, Hamas, and the Houthis, an unusual cluster of sanctioned actors converging on one Russian financial network.

(02:23):

One A7 address received more than sixty-five million directly from an IRGC-attributed address. Another received five million from Hamas. TRM also found that roughly thirty-five billion of A7's reported activity is circular, meaning internal transfers settling on its own books rather than genuine third-party volume.

(02:46):

Crypto inside A7 serves two functions: first, converting USDT to rubles for outside parties and for internal accounting. Separately, A7's ruble-pegged stablecoin, A7A5, helped migrate frozen Garantex customer balances to successor exchange GrinEx after Garantex's twenty twenty-five seizure. Concrete proof that this infrastructure keeps sanctioned money moving even after a takedown.

(03:20):

For compliance teams, treating A7A5 and its affiliated infrastructure as high risk regardless of counterparty is the operative lesson here.

(03:30):

And now, Nick Leiserson.

(03:34):

Nick, thank you so much for joining TRM Talks.

Nick Leiserson (03:36):

Thanks, Ari, for having me. It's great to be here.

Ari Redbord (03:38):

We were on a panel at FINRA, I don't know, a couple of months ago, and literally during the panel I was basically like, "I got to get this guy on TRM Talks because he's so smart."

(03:48):

And quite frankly, this is a topic that I think we're seeing more and more interest in from our listeners, and quite frankly, I'm trying to learn a lot more. So really excited to learn from you today.

Nick Leiserson (03:58):

I'm excited to be here.

Ari Redbord (04:00):

Before we kind of get started, would love to hear a little bit about your journey into the space.

Nick Leiserson (04:04):

Sure. Yeah. I mean, I did not really think I was gonna go into cyber policy. I thought I was gonna be a code monkey. I was trained as a software engineer. I had this budding interest in public policy, but it wasn't until right at the end of my undergrad that I learned about the idea of the Hill, Capitol Hill, this magical place where policy got made.

(04:26):

And I'm like, "You know what? This sounds fun. Maybe I'll go check it out." and when I joined his office in 2010, I was literally on my fourth day in his office when he introduced the original National Cyber Director Act.

(04:40):

So he was already into this cybersecurity thing in 2010. This is Congressman Jim Langevin, just a lion of the field, and he liked having me around as, you know, a 22-year-old, fresh CS grad as a BS detector. Like, that was the number one thing, 'cause people would walk in and say, you know, XYZ claim about what was going on in cyberspace, and then he'd kind of glance over, and I'd either give the, "Yep, okay, that sounds like a real, a real answer," and sometimes it was the thumbs down, and like, "No, these guys are snowing you."

(05:13):

And even in 2010, like, there just was not a lot of expertise on the Hill. A lot of the conversations you hear today about capacity in Congress wasn't there. And me, little old me with just my undergrad, was still one of a literal handful of people on the Hill who actually had experience coding before they showed up. And then I spent a dozen years with him.

(05:37):

He's one of the greatest public servants you'll ever meet, culminating, I've worked with him in every job you can have in a congressional office up to his chief of staff, on the Cyberspace Solarium Commission, and eventually, ten and a half years after I joined his office, the National Cyber Director Act that I ended up inheriting and then became the lead staff author of what we actually got out the door became law in the '21, the fiscal year 2021 National Defense Authorization Act. And basically, as soon as we got it funded, it's another story, it took, another 11 months, I went over to the White House to join Chris Inglis in his front office first to stand up the office and then heading the critical infrastructure and regulatory work over at the White House, which is a lot of fun.

(06:24):

And today, I find myself in civil society getting to still work on these crunchy cybersecurity problems. And what's kept me here, right, as someone, again, who originally thought I was gonna go develop software, not work on policy to help secure it way upstream of that or at a higher strategic level is, it's fascinating.

(06:45):

It's a bipartisan area. It's an area where, you know, it's new policy, it's new national security policy, truly not anything that we've seen before, and that is a very exciting place to be. And I love talking about it, so.

(07:17):

You got the right person.

Ari Redbord (07:18):

I'm super excited to talk about it today. One, one interesting thing, and it's a theme, we've done about 100 and, I don't know, 30 "TRM Talks," and it's a theme that almost goes through so many of them, and that is understanding the technology in order to sort of do the policy, do the legal.

(07:20):

That has been a huge part of your career, is really being able to do this work at a policy level because you deeply understand the tech.

Nick Leiserson (07:22):

Yeah, absolutely. I mean, it is core to what I recommend to people, right? Like, they-- when people come and say, "Nick, you got a CS degree. I just came into this through a more traditional policy area. What, what do you think you should do?" I'm like: Well, take a networking class, because in my networking class in undergrad, our second project, they handed us the RFC, the request for comment from the Internet Engineering Task Force for IP, the Internet Protocol, and our professor was like, "All right, you got, like, three weeks. Go implement IP. Write your own version of the Internet." And there is nothing like that.

Ari Redbord (07:59):

Digging in a little bit. I would start in the past, but instead, I am going to start right now, and that is we are in the midst of a really unprecedented moment in the cyberspace in that we are seeing bad actors in the last few weeks, engage in attacks at unprecedented speed and scale. Would you just walk us through what we are seeing in this moment right now, September, October 2026?

Nick Leiserson (08:24):

Yeah, absolutely. It's, it's a lot. And, it is the culmination, I would say, of a lot of what we've been talking about since about the March timeframe in the broader community. So the first big story that is a good example of this is PHP to Shell. And, you know, there's a lot of reasons that this is a really good incident or vulnerability to unpack. One is that it tells you a lot about people's intuition, going back to that word, right, of people's intuition of how technology works, where if you go to most policymakers and you say, "Hey, did you know that this thing you've never heard of, PHP, powers seventy percent of the websites that are used in the entire internet?"

(09:19):

They're like, "No. I'd never heard of this. How is it possible that this thing that is so vital?" And then you say, "And did you know that there are basically a half a dozen people who maintain this thing that keeps the entire internet, or seventy percent of all websites on the internet going?" They're like, "What are you talking about? This is, that's, that doesn't sound good at all. That doesn't sound very trustworthy. That sounds pretty fragile." And yet that is the reality of what happens all the time in technology, is we place our trust without really examining or really having any intuition as to what's actually underlying these systems that we rely on. The other story, though, with PHP, beyond the broad scope of the impact of a vulnerability that lets you get remote code execution and take over whatever website it is how it came about. And I've seen a lot of people describe these vulnerabilities as being discovered by AI in August of this year.

(10:21):

It's like, no, they weren't discovered, right? PHP, to their credit, the PHP foundation has been incredibly transparent. They're like, "Yeah, we found these in, hmm, late '23, early '24, that timeframe. We found the underlying bug that eventually manifests in October twenty twenty-six as a mess. We found the bug, but we didn't know if it was exploitable, and it was a real humdinger of an engineering problem. And by the way, we were concerned that some of our constituents would be upset by some of the fixes that we had in mind. So we just kinda shelved it, and we're like, "we're gonna leave this to the side. We'll get to it someday, I'm sure. We're tracking it, but it doesn't seem exploitable." And when you pointed a cyber-capable AI model at it in August, it was like, "Oh yeah, it's definitely exploitable.

(11:14):

Here's the exploit, right here. I've got it." Like, I will break your entire stack off of this memory management problem in PHP. There are few better microcosms of where we stand in twenty twenty-six than that. It's like there are problems that we've sort of known about for a while and just relied on the fact that exploiting them was hard, and that's what changed, is no, actually the exploitation is a lot easier than you thought, and that's a big problem because the reason it got put off in the first place was the fixes weren't that easy to come by.

And we're just staring at that all over the place. Couple of the other things that are happening of late, another one that's got me super nervous is a certificate authority, based in India that, had eight certificates that they messed up the signing ceremony for. So instead of having a one hundred sixteen-digit number as the private key, so you can think of that as like the password to get access to this, so it's very long, the passwords for all eight of these were less than, a number, less than two hundred fifty-six.

(12:29):

One of them was four. And this is another area where it's just there's trust in the system that breaks people's intuition because they're like, "Well, you know, what does that mean, Nick? What does it mean that there's this private certificate with a password of four?" And the answer for most people, thankfully, is not a lot.

The browsers have already revoked it. You'll be fine. But if you rely on embedded devices or anything that's ever pinned a certificate. From now till years, like we are years and years going to be dealing with people being able to impersonate literally any website based on this trust. So it's one little failure of, "Whoops, we messed up this key signing ceremony," and the result is if you are a medical device, you might trust basically anything someone tells you, whether it's to intercept traffic or to present you with new code and say, "Oh yeah, I signed this," based on one human error in this case.

Ari Redbord (13:31):

Yeah, no, absolutely extraordinary. How are you seeing AI deployed by cyber actors, you know, over the last year or so?

Nick Leiserson (13:39):

It's a really good question. I was at Singapore International Cyber Week and this was a question that came up all the time, as threat analysts were briefing policymakers at the Counter Ransomware Initiative. So this is year six of this international group of now seventy-six countries, and multi-party agencies like Interpol all getting together to handle ransomware.

(14:08):

And the question on all of the threat analysts' mind is like, where is the AI? And the answer was, we see hints of it, right? We see hints of it. There are penumbras, there are reverberations somewhere, and we're like, we know people are talking about it on dark web forums. We know people are stealing tokens, that they're presumably using for stuff.

But we can't, like, see it manifest in most cases right now in the incident data. And I think that that's actually, a really good where things stand right now, right? Which is the bad guys are experimenting. Some of the areas where they're going to have the most immediate success is not gonna manifest on your network.

(14:52):

It's gonna be, "Oh, I stole five terabytes of data two years ago. That was an impossible task to sort through. I would do a few quick keywords I would hope that I hit upon. Oh, the CEO, well, I could look up their name, and I've got their email. Great. That sounds valuable. I could blackmail against that."

But most of the data you've got, it's like, I don't know how I'm gonna read through five terabytes of random stuff. And now it's like, oh, I can feed that into an open weight model, and it'll spit out here's the important stuff. Like, that doesn't manifest on a victim's network, but it certainly manifests in the ransom note that they get.

And similarly, the reconnaissance and the stuff that they're talking about on the front end of how do I get onto a system. And so the big picture here, right? And what I was hoping people would take away from, the threat briefies we were getting in Singapore is are innovative criminals out there.

(15:50):

So the problem is the trend line is the wrong direction, and you need to keep in mind that, like, by the time you see this manifesting on defenders' networks every day, you are out of your window to do anything about it. So you should be very grateful that we still have enough of a preview of people are still working out exactly what these workflows are going to look like in terms of the core part after reconnaissance and before processing the exfiltrated data of operations on defenders' networks if you're a criminal or a nation state.

Like, the fact that we're working that out is lucky us, but we need to take advantage of that window now, not wait to say, "Oh my God, did you see this AI thing? It's completely crushing everyone's networks." Like, it's too late then.

Ari Redbord (16:43):

Tell me a little bit more about the Counter Ransomware Initiative, the sort of origin story for you. But then also, it must be an incredibly different moment today when you're thinking about that group and this space than it was in the wake of Colonial Pipeline.

Nick Leiserson (16:57):

We at IST ran the Ransomware Task Force, and this was a convening of, private sector and civil society folks. We came out with a report two weeks before Colonial Pipeline. I was still on the Hill, so I.

Ari Redbord (17:12):

That is wild. Okay.

Nick Leiserson (17:13):

So I was still on the Hill. I got the brief from the folks at IST who were doing this excellent work, and then a week later, for me, Colonial hit, and all of a sudden Langevin, who's been talking about ransomware for years, and everyone's like, "What the hell are you talking about, dude?"

(17:30):

They're like, "wait, Mr. Langevin, what should we be doing about this?" And it was very convenient for me to say like, "Hey, we have some ideas that have been kicking around, but there is this group of experts who have really done a bang-up job of laying out 48 separate recommendations of what do." One of them was to convene an international body around this.

And in August of that year, Anne Neuberger, who was the deputy national security advisor for cybersecurity and emerging technology at the time, did just that. She got around 30 countries together into the International Counter Ransomware Initiative to say, "We need to convene around this issue because it's a national security issue."

(18:12):

And as a deputy national security advisor, I recognize that, and saying this is just a law enforcement issue or we can leave it in the realm of let law enforcement and technical folks deal with it, not bring the full national security apparatus to bear, is not sustainable. And Colonial and shutting down 40% of the gasoline supply east of the Mississippi River was like the switch that flipped in many people's minds that said, "Aha, this is a national security issue."

If you fast-forward now and say we're six years into CRI and, or this is the sixth summit, we're five years past Colonial, what does the landscape look like? There's good news and there's bad news, right? The bad news is criminals are very innovative, and thanks in part to, their ability to use new payment technologies, they are finding all sorts of clever ways to continue to threaten and scam people the world over, not just in the US, the world over, and extract an enormous amount of money from folks.

(19:23):

The good news is the quality of some of these hacks is going down, and that's good, right? Like, if you look at Colonial, we did not have the option to turn the systems back on. That was not-- There was one part of the Colonial pipeline that actually ran while this incident was going on.

It was between, Baltimore and BWI, the airport in Baltimore, because there's one customer on that little spur of the pipeline. It was the airport. Every packet of oil that entered in Baltimore into that part of the pipeline went to the airport. So you didn't need any IT to run it 'cause you're like, "There's one customer.

I just need to run this flow. I know how many, you know, gallons per minute are moving." Great. The rest of the system was shut out. There's nothing you could do, and that's why it was a national security issue. Today, and one of the things we talked about at the CRI meeting is, you see extortion-based attacks, but they tend to be based on confidentiality.

(20:24):

That is to say, "I stole your data. I'm gonna publish it on a leak site unless you pay me a ransom." And that's not good. But it is better for my money than saying you can't get access to your systems. And I think it's important for policymakers and folks in this community to look at, like, well, where are we succeeding in driving away from availability attacks and into confidentiality attacks?

'Cause that's a better class of problem. Still a problem, but it's a better class of problem. And if part of the reason, for instance, that if you as a victim feel the need to pay a ransom is because you're afraid of lawsuits about a breach of data, well, can we do something on that side of things which doesn't look the same as when it's availability, and I just can't turn the pipes back on?

Ari Redbord (21:12):

I'm being a little glib here, but like other than negotiating our way out of this with ransomware groups, right? You know, I think we've gotten a little bit into sort of like how bad actors are at least experimenting with AI to supercharge these types of operations, right?

Targeted phishing attacks, deepfakes, much more sophisticated cyberattacks. But moving a little bit to sort of the solution piece. You know, I've been arguing for some time that it's critical that national security agencies, defenders, right, law enforcement have the same access to tools, that they're able to use tools to, use AI to run over systems in order to identify vulnerabilities.

But how do you think about the solutions here as you're talking about the problems?

Nick Leiserson (21:50):

The first most important thing is we have to be in this together, right? Like, that is core to our success in this space as defenders is the bad guys look for seams. They look for gaps in our broader trust ecosystem. That's what they exploit.

Chris Inglis said it much more elegantly than I ever could, which is, you know, what we are aiming for in this space more than anything is to create a situation where you have to beat all of us to beat one of us. That this is a team sport, and that team sport extends from private industry to government.

Like, if we are not working together in this space, we will lose. So that is foundational to everything else that I'm gonna say in terms of what do solutions look like. Some of them look like pushing things upstream, right? That, that was a major focus of the strategy that we worked on in the last administration, and it remains true.

(22:54):

There are places of higher leverage in the ecosystem, whether it is software and taking advantage of the amazing tools that are finding vulnerabilities to say, "Hey, we can actually reduce vulnerability density. We can use technology to make more secure, higher quality software for the first time." Those kinds of improvements redound across the entire ecosystem.

Everyone benefits when your software quality is going up. Or with service providers and saying there's more that service providers can do to protect their customers, and that in a shared responsibility model where it's customers and it's providers that have some responsibility for their security, but the arrow needs to go towards more responsibility on the providers.

(23:43):

And that remains true as well in an era of who's going to be able to take advantage of new technologies like AI first. Then there is the how do you take advantage of new data streams that-- well, new but better ability to process data streams. And I will tell you, like, the first times I was getting briefed in twenty sixteen and seventeen on, like, what does blockchain intelligence look like?

Like, what do analytics look like? I said, "Wow, this is going to be really useful for cyber defenders." Separate and apart from, like, you can get tactical information about what are tactics, techniques, and procedures that carry across from group to group by understanding how the money flows in the background of that.

(24:37):

And that still, in twenty twenty-six, isn't done as much as I would have thought at the time, and it's because it's hard. Like, it's hard to map this stuff. But there are technological changes that are making that easier. And so just as the bad guys are getting better at taking a blob of data that they've stolen and figuring out, "This is what I want to extort from," we need to do a better job of saying, "Hey, we've got intelligence across all of these different sensors, including non-traditional ones," and fusing them in to a more complete picture of like, oh, actually, we know something about how to defend our networks and our systems because we know something that looks orthogonal about how the money is moving over here.

And drawing those connections was important ten years ago. It's just as important today, but we keep seeing better backstopping from a technological standpoint in our ability to do so. So that's another thing that I think, you know, you look at what does defense look like and what gives me hope. If we can be as nimble as the criminals in terms of saying, "We need to take advantage of these new opportunities," we can make progress.

(25:54):

We've done it before. We'll do it again.

Ari Redbord (25:56):

Any sort of cool examples that resonate for you in terms of what you just described?

Nick Leiserson (25:59):

Yeah, I think that you've seen really just this year, right, a fair uptick from the Bureau continuing to do disruption operations and prioritizing that in their cyber strategy. And that runs the full gamut of what they're doing from residential proxy networks, which are just a completely underdeveloped story of where bad things are gonna be coming from for the next several years, to, you know, Russian networks that have been used by the GRU as obfuscation relays as well.

But even then, you know, I think that the next, the next level of how do you use this intelligence is to say, yeah, you know, first order is follow the money so we can get it back and we can say, "Hey, the criminals want to crime because they can make money out of it. If we could take the money out, then they're not gonna be as interested."

(26:55):

But the next level is to say, hey, knowing that all of these seemingly disparate, affiliates are actually using the same core infrastructure, ransomware affiliates are using the same core infrastructure. And the way that I knew that is that I followed the stuff that went from a ransom payment into, you know, an affiliate's wallet, and then I tracked that over to the actual service provider that's providing the backbone for all of this and saying, oh, with that, I can, as a defender, protect myself because I know that's actually the target.

That's what I should be looking at. That's how I should be dealing with what the threat environment is, not the little fish on... In the affiliates who are coming after me directly, and simultaneously feeding that into disruption operations and saying, hey, regardless of whether it's happening through law enforcement or with what service providers can do.

(28:01):

Like we're right, working on a paper right now taking anecdotes of what service providers can do with no additional legal authority, right? Not with a court order, not with a warrant, nothing else. What they can and have done to take disruptive action just when they have the right incentives, and what that looks like from a legal standpoint.

But the fuel to those engines, right? If we say like, okay, we've using CISA 2015, the Cybersecurity and Information Sharing Act of 2015, and the broad discretion you've got to monitor there, and the Wiretap Act and the exceptions in there for service providers, and a couple other things, you know, just common law defenses tied to protecting your own property you're going to take action to disrupt an adversary.

You have already all of the authority you need. You just need to be motivated or have the appropriate risk appetite to do so. That's great, but it's still going to need intelligence from a whole host of folks to say, "Okay, but what am I going to take action against?"

Ari Redbord (29:08):

What is your wish list right now, as we're seeing cyber activity scale with AI? What do we need from policymakers on the defense side?

Nick Leiserson (29:17):

What a great question. How long you got?

Ari Redbord (29:20):

What do you want? This Christmas shopping.

Nick Leiserson (29:23):

I think that the first thing that I want from policymakers is to understand where we are in this technological evolution because different players are going to have a different role to play. So like Congress, you know, picking on Congress since I spent a lot of my career there, I want Congress to be looking and saying on the software quality issue, how do we ensure the incentives are such so that when the marginal cost of software falls to almost zero, which it's already almost there, but the cost of securing that software is not almost zero, that the incentives are in place so that we are still actually going to secure the code that's being produced by AI agents.

That is not a today problem. It's a three years from now problem once we get over the hump of dealing with the legacy software. But if we wake up in three years and say, "Oh man, you know, we need-- we really need to change this incentive structure some, and, like, we need some changes to the law to do that," having worked in Congress, you're not going to get that on the timeline that you want.

(30:33):

Like, you need to do that today in Congress. That's where you need to skate to where the puck is going, you know. Do the Gretzky. Don't sit here and chase after the thing that is operationally relevant right now, which is where, you know, back to how do we address the software quality issue. Okay, clearinghouses, how do we make sure that the PHP Foundation, we go find any other skeletons in their closet, every other bug that looks like that?

(We're going to go exhume that right now and make sure they have the engineering resources they need to fix that. That's a tomorrow problem or a yesterday problem. It's not a three years from now problem. I don't need Congress focusing on that. I need the White House focusing on that and CISA focusing on that.

(31:16):

And that's in the software quality sort of sphere. The other side of this is AI as a, capable offensive operator in its own right. Right? Like the hugging face, oh, can AI agents just run end-to-end, cyber operations? And then how do they, in a jagged way, affect different parts the kill chain? And what I need policymakers to be doing is saying, "Okay, in that world, what kind of investments do we need for defense, and how do we make sure those scale?" And again, strategically, I would say focus on pushing them upstream and not saying to everyone who's gonna exist on the internet, "You better figure out your own AI-powered defense."

(32:07):

But those are the conversations that need to be happening right now with policymakers.

Ari Redbord (32:14):

It feels to me like things are going to get worse before they're going to get better. We're just not ready, right? Bad actors are moving faster than we are, which is, kind of historically true, right?

(32:27):

Automobiles, end-to-end encrypted messaging apps, crypto, now AI. We're building defenses, but it's inevitably slower. Are we going to continue to see this really unprecedented scaling of attacks, numbers, and sophistication?

Nick Leiserson (32:42):

The odds are definitely better that we'll continue to see that, right? And I think really depends on your market segment exactly how sure I am. If you are in, you know, what we call the target-rich cyber poor, if you are an under-resourced critical infrastructure operator, absolutely.

(33:03):

That I can say unequivocally. And a lot of those folks are like, that's-- When I say it as an under-resourced critical infrastructure operator, people are like, "Well, what does that mean?" I'm like, "Okay, it means your school," right? It means where your kids are going to school right now. It means your water utility.

(33:20):

It means your hospital. Exactly. And, like, they are going to suffer 100%. There's, there's very little we can do about that as this problem scales, and they urgently need policy help. The race at the more top end is a little less clear, right? Because again, there's one way to look at PHP and say, "What a mess," and there's another way to say, "We did get a head start on the, on the bad right?

(33:47):

Like, we did with AI tooling, we were able to say, "Yep, the exploitability is definitely there. Get a patch in place," and beat out this being, oh, it's O day or worse, the PLA has been quietly, or the Ministry of State Security has been quietly, sort of exploiting websites like this for years before we figured it out, and instead we found out and the bad guys found out at about the same time.

(34:24):

That's not ideal. Ideally, we beat them by a lot, not by, "Oh my God, patch now," right? The Prime Minister of Japan sending a tweet to the whole country at 6:00 AM saying like, "If you own a web server, please patch today." like, not on my bingo card of what I expect to see. So if you look at it through the lens of will we see PHP-like incidents go up, like yeah, definitely. It's still a little early to say exactly how bad that is other than exposing what it is demonstrating is, oh man, we've built on a house of sand, right? Like that, this is, we've built a whole lot on a foundation that's not too steady, but we are at least racing some to catch up there.

Ari Redbord (35:10):

There is no way I can end this on civilization-level threats to hospitals, water filtration, and electric grids. Nick, what do you do when you are not trying to stop attacks on US critical infrastructure?

Nick Leiserson (35:23):

So of late, a lot of it is baking with my four-year-old daughter. So this weekend, you know, we just made yesterday an apple cake with some farmer's market apples and a bunch of banana bread. But she's into whatever we're baking this week is like, she's like, "Okay, what are, what are we doing?"

(35:43):

So that is, you.

Ari Redbord (35:44):

I love that. Are you guys good? Have you made anything that you're like, "Wow, this is amazing"?

Nick Leiserson (35:49):

I mean, I'm pretty-- She, she's-- Someone asked me today, they're like, "Is she net, is she net chaos?" And I'm like, "She's about net neutral now." So good for you, Amelia. Like, as a four-year-old, it's pretty, it's pretty good to be net neutral. I think we'll get to value add, but she's like, "I've got my whisk skills way.

Ari Redbord (36:09):

I think that's really cool. It's also like this whole idea of like making things, right? Like of putting the ingredients together and coming up with something better than what you started with. I love that.

Nick Leiserson (36:17):

Yeah. And it's super not like, it's one of things that you can.

Ari Redbord (36:23):

No screens.

Nick Leiserson (36:24):

Cooking is like, yeah, it requires just the right level of attention of, you know, you gotta pay it, you gotta pay some attention, you gotta use 30% of your brain. You cannot ignore your banana bread in the oven or your, you know, stew on the stove or whatever else it is.

(36:43):

But it also does not require, like, complete engagement in a way that when you're thinking about, you know, what exactly does this mean for hospitals? Like, no, I better be bringing my full attention to this.

Ari Redbord (36:55):

I love it. Nick, thank you so much for joining TRM Talks.

Nick Leiserson (36:58):

I appreciate you having me, and thanks for uplifting the issue.

Ari Redbord (37:05):

You know, it's interesting. I joined TRM not all that long before Colonial Pipeline, which was a really-- I think it was a real watershed for the world when it comes to ransomware and cyber. It was a watershed for me. I'll never forget, you know, being on calls with the FBI, more or less for the first time, I think, as a private company, thinking through how we can attack, track, and trace the ransom payment in that case.

(37:29):

But it's amazing how far things have gone and how fast they've moved. And I think that having this conversation, in this moment, which feels to me like another watershed. I heard someone refer to October 2026 as Hacktober. We are seeing unprecedented scale when it comes to these attacks because of AI.

(37:48):

And, you know, two weeks ago or three weeks ago, we announced Ardence, which is the defense. It's how we can combine on and off-chain data in order to deploy that type of data across systems in order to identify malware, to patch vulnerabilities. So what an incredible moment to have this conversation on really understanding the threat landscape that Nick is so expert on, but then also loved hearing from him on what we can do potentially from a policy perspective, what we can do from an operational perspective.

(38:18):

So, I said at the end, I learned so much today, and I hope all of you do too.

(38:23):

On the next TRM Talks, I sit down with the Global Head of AML for PayPal, Kate Salottolo.

(38:33):

If you love the show, leave a review wherever you're listening to it, and follow us on LinkedIn to get the latest news on crypto regulation, compliance, and investigations.

TRM Labs (38:44):

TRM Talks is brought to you by TRM Labs, the leading provider of blockchain intelligence and anti money laundering software. This episode was produced in partnership with Voltage Productions. The music for this show was provided by iKOLIKS.

Ari Redbord (39:00):

Now let's get back to building.

About the guests

Nick Leiserson
Institute for Security and Technology (IST)

Nicholas Leiserson is the President of the Center for Advancing Cybersecurity at the Institute for Security and Technology (IST), where he previously served as the SVP for Policy. A legislative strategist and technologist, he has spent 15 years addressing cybersecurity risk and resilience and managing multidisciplinary teams of senior professionals at the White House and on Capitol Hill. He was integral in the creation of the Office of the National Cyber Director (ONCD) and served as one of its first employees.

Throughout his career, Leiserson has focused on developing policy ideas, building coalitions around them, carrying them through to execution, and conducting oversight of implementation. Leiserson previously served as the Assistant National Cyber Director for Cyber Policy and Programs at ONCD. He was recognized for his work leading the creation of the National Cybersecurity Strategy Implementation Plan and developing the Biden Administration’s approach to cybersecurity regulatory harmonization. He earlier served as ONCD’s Deputy Chief of Staff where he built internal policies for the first new component in the Executive Office of the President in three decades and oversaw operations, including staffing and budgeting.

Prior to joining ONCD, Leiserson spent more than a decade on the staff of former Congressman James R. Langevin (RI-02) in various roles, most recently as his Chief of Staff. He created and executed the House legislative strategy for the Cyberspace Solarium Commission, which saw more than two dozen of the Commission’s recommendations enacted within nine months of its report being issued. He holds a degree in computer science from Brown University.

More TRM Talks

EP.   |  Nov 11, 2026 - 26min

Inside the $6.36B Binance Hack with Nick Carlsen
PLAY EPISODE

EP. 123  |  Nov 4, 2026 - 39min

Every Lawyer Is a Technology Lawyer Now: Lewis Rinaudo Cohen on Machine Law
PLAY EPISODE

EP. 122  |  Oct 21, 2026 - 30min

Building Ardence: Combating Modern Crime On-Chain and Off-Chain with AI
PLAY EPISODE

EP. 121  |  Oct 7, 2026 - 36min

Q Day Is Closer Than You Think: Quantum, Crypto, and National Security with Sanjeev Bhasker
PLAY EPISODE

EP. 120  |  Sep 23, 2026 - 36min

From Emerging Markets to Tokenization with Morgan Stanley’s Amy Oldenburg
PLAY EPISODE

EP. 119  |  Sep 9, 2026 - 40min

How Institutions Can Use Permissionless Rails with Rebecca Rettig and Michael Mosier
PLAY EPISODE

EP. 118  |  Aug 26, 2026 - 31mins

Inside Australia’s Historic AML Reform with AUSTRAC’s Brad Brown
PLAY EPISODE

EP. 117  |  Aug 12, 2026 - 30min

A 1934 Footnote Predicted Governance Tokens: Rethinking Crypto Law with Teresa Goody Guillén
PLAY EPISODE

EP. 116  |  Jul 29, 2026 - 36mins

Treasury’s War: Then and Now — Economic Statecraft in the Age of Technology with Juan Zarate
PLAY EPISODE

EP. 115  |  Jul 15, 2026 - 33mins

Criminals Don't Have Compliance Officers: The AI Arms Race in AML with ComplyAdvantage’s Vatsa Narasimha
PLAY EPISODE

Subscribe to TRM Talks

Subscribe to be the first to hear about new episodes, and to stay in the know about all things blockchain technology and crypto policy.