HSDF THE PODCAST

From Innovation to Action Integrating Technology in Law Enforcement - Part 2

• Homeland Security & Defense Forum

Share episode

Integrating technology in law enforcement transforms innovation into actionable tools, enhancing capabilities in crime prevention, investigation, and community engagement.

Exploring the intersection of technology and human factors in security, this episode addresses pressing challenges in cybersecurity and operational efficiency. It emphasizes the critical nature of training individuals to recognize threats and advocates for the integration of innovative technologies to enhance decision-making processes.

  • Michael Prado, Deputy Assistant Director, Homeland Security Investigations
  • Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP
  • Martina Melliand, Director of the Enterprise Analytics Division, CBP
  • Bogdan Frusina, Founder, Dejero
  • Luke McCormack, former DHS CIO (moderator)

This discussion took place at the HSDF’s Border Security Symposium on December 11th, 2024 

Follow HSDF THE PODCAST and never miss latest insider talk on government technology, innovation, and security. Visit the HSDF YouTube channel to view hours of insightful policy discussion. For more information about the Homeland Security & Defense Forum (HSDF), visit hsdf.org.

• Luke McCormack, former DHS CIO (moderator):

Fair enough, we're going to take two more questions here at the panel and then open it up here. So comms, secure comms right in your wheelhouse. She was describing that. You know what does it take to pull that together to make sure that you have that done right and gain the trust of the operator when that gets delivered?

• Bogdan Frusina, Founder, Dejero:

Well, there's a couple of real world here I know, I know it's it's so practicality it is.

• Bogdan Frusina, Founder, Dejero:

I don't think there's such thing as 100% secure system in this world, because there's always a human factor which always causes a problem as far as security goes. There's too many phishing, and so I think the most important thing about security, whenever somebody asks me that question, is about training the humans that are behind it and making sure they know how to protect their own information so they don't get scammed to enter the system. Systems themselves, they're actually quite well designed in a way that you can't easily hack them. We had a conversation earlier, actually, with somebody about quantum computing and that may be happening in the future. There might be a risk, but at this point in time that doesn't exist yet as a risk factor. However, making sure that you follow the standards you see all the time all these security bulletins that come out with all the holes that exist in security infrastructure so making sure you're up to date on that it's extremely important and, look, it is a nightmare out there as far as the amount of holes that exist and the amount of ways around things, but I firmly believe that if you follow a proper standard.

• Bogdan Frusina, Founder, Dejero:

I think I was speaking with Mr Bhagavalia a while back and he told me something like 180 million attacks a day that CPP handles per day. That's how many attempts there are on the websites and all the infrastructure that they've got. It was a crazy amount. It got stuck in my head. As far as the number goes, it's absolutely ridiculous.

• Bogdan Frusina, Founder, Dejero:

So, obviously, making sure that you have all the nearest vulnerability and all that stuff, but the only way you can do this is make sure you're really quick at patching, you're really quick at being aware of notifications and it comes back to a very simple thing I was going to mention earlier, and the helicopter example that you've done is, I think the most important thing out of this entire conversation is how do you as was discussed is like, how do you get all the data put together, but how do you create actionable intelligence, which is what this is that delivers information that's pertinent to that particular information for that particular officer in the field, to ensure that they do their job in the best way possible and they're aware of the most amount of information possible. So the UI and the interface is probably equally important to the good of the data below that.

• Luke McCormack, former DHS CIO (moderator):

Customer experience. We could talk all day about that. I want to talk about speed. Go back to B-Pod and talk about I don't know. I have a sneaking suspicion that one's going to look at this way, where you're acquiring goods and services, and say, boy, why aren't we moving at the speed of light? Why aren't we? You know, what are these CSOPs? What are these OTAs? What are these SIBRs? Why don't we do things faster than we're doing them today? I know you all have done some of that. You want to speak to sort of non-FAR versus traditional FAR type acquisitions. Any thoughts on that?

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

Sure, and just to make it clear, I don't represent the Border Patrol Program Management Office Directorate. I've worked with them for years. They're great partners.

• Luke McCormack, former DHS CIO (moderator):

So don't pay attention to the uniform he's wearing. Anything I say that's correct is because of that.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

Anything that's wrong is on me.

• Luke McCormack, former DHS CIO (moderator):

Yeah.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

Well, I mean so. So look, the far was put in place to ensure fair and open competition and compliance, make sure that the government can get access to the best goods and services, um, you know, in a fair and equitable way. However, it's still very social. It's a social process. A lot of the things that we discussed today are social problems. People have to talk to other people, write down words for other people to vet, for other people to read, for other people to understand, to then have meetings to be understood, to then go back to me on a panel to judge the words people wrote, and that just takes so much time.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

And back to AI, that we should be focusing, in my opinion, on using AI tools to help us accelerate our business processes.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

Ai tools to help us accelerate our business processes.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

I mean, there is no reason why we can't have a data sharing policy, compliance ontology that would be able.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

We know the policies and laws, we know the authorities, we know the data elements, and so we know the rules in the FAR that it should be able to recognize.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

I should have an ask and tell system that says I want to accomplish this contract for these resources, these parameters, using this data, and then you would be able to quickly pull the laws that govern it in the FAR and in CRCL civil rights, civil liberties, privacy, ptas, pias, all of that and then pre-build your research packet that you know you have to do every single time over and over and over again. Those are the things that are appropriate for AIML tools to routinize the things humans are always going to have to do with the FAR or non-FAR OTA mechanisms. So that's what I think is really an opportunity is what are the ways we can accelerate our business processes by automating the human functions to fill out all the different paperwork we have to, which is important for policy compliance, privacy, fairness and open competition, but do it faster so that humans only have to deal with the judgments that require human on-the-loop decision-making. Give me the research packet ahead of time that accelerates all the paperwork you must have.

• Luke McCormack, former DHS CIO (moderator):

And I don't think anyone's going to argue with that out there. Any questions? Got one right here. So great comment on data. Data is so important. So, angelique Irving Fairlight, try to elevate it up there, because we don't have a Angelique Irving Fairlight. We make the surveillance systems on the northern and southern borders. We're large providers. You got a mic coming right behind you there. Whatever percentage.

Audience:

Here you go. If half of those systems are down or whatever percentage, how do you get good data? And if you don't have repair budgets but you only get new system budgets, how do we continue that operational piece and collate all that data in a way that helps our mission? For the guys in the field who don't have eyes, because they need them, and I think that's my question, how?

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

do we deal with that? I mean that's terrible. I had to deal with it when I was division chief in Buffalo sector over the technology portfolio and things would go down and parts would be on back order and you have to reprioritize and you have to take folks that were doing other things that were more productive and covered down in your areas of high threat and high interest. So you go back to basics and at the end of the day, board patrol is about looking at the ground and tracks and driving down the road, pulling a tire and looking for data, whether it's with the agent every agent's a sensor or the technology. So when technology goes down, it's a massive hit to our productivity and efficiency.

• Luke McCormack, former DHS CIO (moderator):

Next question.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

Sorry, I jumped in, that hit me.

Audience:

Ryan, congrats. Entertaining as always. You know to kind of take it at a larger perspective, at the integration of the tech. You know we've got three fairly distinct roles here HSI, cbp and Border Patrol. What are the common threads you see when you look at bringing in a new tech and that opportunity for the force multiplier, reducing the cognitive load, or new sensor phenomenology, where do you see the greatest pitfalls or the most common pitfalls and where do you see, you know, the sort of the best model to follow? What's the best practice?

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

That's a fantastic question. I always start with going to what we can all agree on, and we can always agree that we want our agents, officers, special investigators, pilots, analysts, whomever to be safe and effective. Then, second, maybe the pitfall is we all have to move out on our mission and we all get judged by our boss or our performance plans on what we're supposed to do and sometimes we get tunnel vision on making sure our program is successful. And often it's useful to remind ourselves to step back and say how am I making the mission successful, making sure those align? And at the end of the day, back to the social problem. You can't automate collaboration and relationship building. So there's always a person aspect of me going to my partners in HSI, cbp, industry and say, hey, this is what I think I need and this is what I want to try to do, and then make sure that we have that shared understanding or shared effective understanding, and it's people, at the end of the day, that at the beginning and end of that loop.

• Luke McCormack, former DHS CIO (moderator):

Any other thoughts on that one, any other questions? Because I got a couple of them.

• Luke McCormack, former DHS CIO (moderator):

I got a whopper here to close it out, one more here and then I Thanks, good discussion, just curious if you could speak to some of the laws and policies if there's any obstacles for integrating more tech. I know some of the other countries and state and local law enforcement have different laws and policies that sometimes make buying and integrating things easier.

• Michael Prado, Deputy Assistant Director, Homeland Security Investigations:

Thanks, yeah, that's a fantastic question. I actually was discussing this this morning, that exact issue. I'll use the generative AI and some of the AI tools that HSI specifically for child exploitation investigations, investigations of networks of individuals who are sexually exploiting children, online posting those images, trading those images, some of the tools that we use. While extremely effective and extremely powerful, because we're a transnational criminal investigative agency, we have to work and rightfully so should be working with our foreign partners. Some of our foreign partners, such as those within the EU, take a wholly different perspective on the use of certain AI tools for investigative purposes because their privacy thresholds are so much higher in certain areas within the EU construct. That poses an issue.

• Michael Prado, Deputy Assistant Director, Homeland Security Investigations:

So, being able to work side by side with our foreign partners and I think I mentioned earlier we have 90 foreign offices, so, by definition, most of our investigations are transnational in nature we have to take into account what we can share with our foreign partners, what kind of cases we can work. We work child exploitation cases with them, but we have to take extra care in the development of those investigative leads, how we come to share information. Things along those lines make it complicated and I anticipate and forecast that those issues are going to just get more complicated as we continue to push forward in the coming years, certainly with additional generative AI and AI-powered tools, large language models and things of that nature. That may give some of our foreign partners, who are absolutely critical in our global investigations, a bit of pause when working with us. So we've got to work that out on a policy level.

• Luke McCormack, former DHS CIO (moderator):

I'm going to wrap it up with one last question. For each of the panel members, same question. Let's talk about sort of the next biggest threat, cbp. I want to start with you. And sort of fast forward, we have a lot of knowns out there. No doubt there's some unknowns. Fast forward, we have a lot of knowns out there. No doubt there's some unknowns.

• Martina Melliand, Director of the Enterprise Analytics Division, CBP:

What's in your job? Jar going forward as far as the biggest threat. This might sound a little revolutionary but, luke, I don't think it matters what the next threat is. It doesn't matter because we aren't effectively utilizing our data right now for our current threats. We have a vast amount of unused data, and I would say that the biggest threat to the CBP mission, and probably even expanding that across the Homeland Security enterprise, would be to say it's our lack of capability for an interoperable data infrastructure that will help us support decision-making across domains. I would wager that to what you know.

• Martina Melliand, Director of the Enterprise Analytics Division, CBP:

Again, some of the panelists have already mentioned on this panel and previously that the next threat we know about it, it is in our systems, but we are not able to connect those dots. Just DHS in and of itself, of which CBP is a part. We have over 900 largely disconnected systems, estimating that about 60% of the data in those systems is unknown or unused. That is just mind-boggling when you think about that. And so we have this mountain of data available at our disposal for exploitation and we're not able to exploit it.

• Martina Melliand, Director of the Enterprise Analytics Division, CBP:

The US government's inability to connect the dots and pull the meaning from our databases that's not new language that we've been talking about today, is it? I would wager that every person in this room, or near everybody, read the 9-11 Commission report when it came out 20 years ago and it said something similar how the US government has a vast amount of data at its disposal and when you don't take into account traditional intelligence databases and you put in on top of that customs and immigration data, the storehouse is immense. But what we as the government are not good at doing is utilizing and processing that information and across the government agencies, official after official, implored the 9-11 Commission to please bring attention to the unglamorous back office side of government operations unglamorous back office side of government operations. And don't get me wrong, a lot of great work has been done in these past almost 25 years since then. But I would ask each of us are we where you think we should be 25 years after 9-11? I think we could be better, we need to be better, so let's be better.

• Luke McCormack, former DHS CIO (moderator):

Some powerful statements there. Border Patrol, I saw you nodding your head there, in agreement.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

Yes, I'm going to take the opposite approach, but I'm going to blame somebody else. A few months ago, former CEO Eric Schmidt of Google said he believes we're going to be in a war for knowledge supremacy with near peer competitors in our time. I'll tell you, I believe that's happening now, and what knowledge supremacy means is the ability to, we say, things like actionable information or data fusion at the speed of the mission. It's exactly that to use our technology and tools to recognize data, convert it into information, add context and meaning, get that only the things humans need to know, or even actuate, kill web, kill chain. From a military perspective, ai versus AI battlefield management that's probably the greatest threat that I see, and how that will happen is the first global power to combine machine learning and machine reasoning. Machine learning is our math. Machine reasoning is the first global power to combine machine learning and machine reasoning. Machine learning is our math and machine reasoning is the graph. And when you're able to put those together to create actionable data in near real time, you're going to have AI versus AI knowledge supremacy, and whoever does it better, with more accurate context and finely tuned models and ontologies, is going to win.

• Ryan Riccucci, Division Chief, Enforcement Technology & Operational Programs, CBP:

And, honestly, that does keep you up at night, because this is all I do. You probably wonder why there's a board patrol agent in a green uniform talking Dennis Miller. I had to look that up, by the way. I didn't understand the Dennis Miller reference earlier, but Dennis Miller, on Monday night football pulls out big words once in a while. Thanks for that, but that's what I think is the biggest threat. Thank you for that. That's kind. It's the country that develops and combines existing technology in a new way and maybe with additional technology like quantum computing to compute at speed and scale, is going to have a massive competitive advantage in all aspects and domains of global society.

• Bogdan Frusina, Founder, Dejero:

DeGiro, your perspective. Well, cyber threats and everything else will arise every single day. That's not the challenge. I think the challenge is the mission that's being talked about here, which is creating a data layer that's ubiquitous and it's able to be interoperable and interactive. From our perspective, the only way that happens is with having a ubiquitous network infrastructure below that, which also doesn't exist, because you've got all these diversity of technologies that are out there, whether it's LMR, which has its benefits, you've got LIOs, you've got GEOs, you've got MIOs, you've got multiple cellular technologies, you've got landlines, you've got talk about all of it and it's not standardized, not interoperable, very difficult to integrate different speed amounts and it literally limits the amount of data that can be transferred, or not, depending on what you're using. So, from our perspective and our goal, what we're trying to do is we're trying to create this network fabric that's ubiquitous and transparent and always on and always available, so that it enables the higher level technology, such as data interaction and data interoperability, in a seamless way. That's how we look at it.

• Luke McCormack, former DHS CIO (moderator):

Mike, take us on Largest threat.

• Michael Prado, Deputy Assistant Director, Homeland Security Investigations:

Yeah, and look, I'm in agreement with everybody on this panel. But certainly where I sit at the DHS Cybercrime Center and I've had a front row seat this is my second tour there now. As I lead it, I see the biggest threat. Look, the threats out there are growing exponentially and it depends on what you define as a threat from a criminal justice perspective, right, dtos multi-billion dollar profit. Profitability by these DTOs are larger than some of our largest corporations and DTOs drug trafficking organizations. Thank you, you know. Obviously a threat China and other geopolitical concerns that exist. But what I'm looking at when I look at the next looming threat, it really it's not just on the horizon, it's already here. It's those potential adversaries again, whether it's a criminal network, transnational criminal network or a foreign adversary.

• Michael Prado, Deputy Assistant Director, Homeland Security Investigations:

Harnessing the technology that currently exists and is available, readily available and is on the horizon, and using that in a manner that helps them obfuscate their activities makes us in some cases blind, as was brought up earlier, right, our continued reliance on an aging infrastructure here in the United States that leaves us open to exploitation by criminal networks, extortion and blackmail, or by a foreign power makes things tremendously concerning from my perspective.

• Michael Prado, Deputy Assistant Director, Homeland Security Investigations:

So the use of these tools, these tremendously powerful tools that we all essentially hold in our hands now. I mean the fact that AI is now being layered into basic commercial software, pre-packaged, without any necessary you know training or anything like that that you know it used to be. We had to worry about a very, very small group of individuals with a lot of knowledge and a lot of abilities. Now pretty much anybody can quickly learn through open source AI on how to create a deep fake and cause all kinds of mayhem in our democratic processes, everything else. So the threats are all over the place. To me, the threat it can be bucketed in simple terms that the use, the misuse of the technologies that are continuously being developed out there as law enforcement and that's my little narrow scope of the world, obviously is how can we as law enforcement, federal law enforcement harness those and utilize those to leverage against the adversaries that are using it against us?

• Luke McCormack, former DHS CIO (moderator):

On behalf of every single American, I want to thank everybody on this panel for everything that you do every day to keep this country safe. Can we have a round of applause for these patriots? Thank you very much.