Skip to Main Content

AI Helps Identify Data Gaps, Improve Interoperability at DHS

ICE and NCITE, a DHS S&T COE, highlighted the ways AI can support DHS mission integrity.

7m read
Written by:
Army soldier using computer for identification blue digital fingerprint
Photo Credit: South_agency

Artificial intelligence use cases to support agency missions can be innumerable. For one, AI can create the bridge between making sense of the data and filling in data gaps.

“Terrorists like to talk a lot online, but don’t volunteer to participate in our studies,” said Gina Ligon, director of National Counterterrorism, Innovation, Technology, and Education (NCITE), during a panel on AI and mission integrity. “We do have a lot of missing data. Fortunately AI helps us to find patterns in that missing data and project what that missing data value should be. That’s been really helpful for us to understand the parameters.”

Ligon’s organization, a center of excellence commissioned by the Department of Homeland Security Science & Technology Directorate, studies domestic terrorism to develop research and technological tools to support the DHS mission.

“We don’t have a formal domestic terrorism charge associated with domestic terrorism acts [so] we have missing values in that,” she said. “Where I’ve found AI to be helpful is to look at those 5,000 cases we have and try to predict what the recidivism rate is for different types of extremists. That’s a particular way we’re using it right now in the center.”

Immigration and Customs Enforcement, and DHS at large, is working hard to establish common data standards and data interoperability to get the most out of AI.

“With AI, having proper data accessibility is critical,” said ICE Chief Data Officer Ken Clark during the panel. “We have a very close relationship with the other components. We have a data governance council, I’m a member as CDO. That gives us a lot of opportunities to do collaboration, coordination. At ICE, we’re developing a data analytics framework to identify the opportunities and technologies — we rely heavily on [NCITE]. We see a connection with academia as being highly valuable to work a number of these solutions we’re looking for. The department has an AI strategy that helps guide a number of efforts and synchronize activities.”

Adan Vela, an industrial systems engineering professor at the University of Central Florida and a researcher at NCITE, sees AI as an opportunity to fill DHS-specific data and research gaps.

“That’s probably one of the best uses of AI — getting an understanding of what’s missing,” he said during the panel. “A lot of the time we think about AI as giving us information, but think of it backward: it’s showing us where the gaps are and giving us tools for investigators to move forward on. How do we manage access to these systems and the data that supports them? I’m particularly interested in how we can use blockchain as part of identity management systems so we can understand who is allowed to access those systems and understand the impact of what each touch is, if we can track who is touching systems and what they’re doing. I think that adds a lot of security to the IT infrastructure that goes behind the AI systems.”

To support data interoperability at DHS, the department’s components are categorizing data into “domains.” This allows components to share best practices for managing data in those different domains and identify new tools and technologies, like AI, to manage data.

“I’m working with the other partner components on a law enforcement domain as it relates to data and synchronizing the needs and opportunities for using AI across the other nine component partners I have working with in the law enforcement area,” Clark said. “We’re also looking at common guidance for the use of commercial data and consistent standardized use. As you develop one tool, that tool could have applicability for another component. We have app stores and ways to access various applications ourselves on personal devices, but how can we get some of these tools out to other components that are successful at ICE? We see these efforts as improving coordination and interoperability across the department.”

ICE recently launched a natural language processing pilot to identify data on a recurring basis to support White House and Congressional requests for information on ICE activities, Clark added.

As ICE continues to brainstorm new ways to manage data with AI, NCITE works behind the scenes to develop tools that add value to the DHS mission and are compatible with evolving IT infrastructure and IT modernization efforts.

“[Artificial intelligence’s] true value is going to come in two areas: automating things that are distracting to the operator or helping the operator gain clarity,” Vela said. “If it can’t do those things, its benefit is increasingly limited. Retraining is a fundamental part of our workforce development. We need to have infrastructures in place that allow us to retrain AI systems. This is all about IT infrastructure and enterprise systems behind AI. I get a little disappointed when people talk about AI and big data, but never talk about the engine behind it, which is IT infrastructure. If you don’t have that, then we can’t talk about retraining or improving our systems, we can’t even talk about adapting as the DHS mission changes.”

Woman typing at computer

Stay in the know

Subscribe now to receive our curated newsletters

Related Content
Woman typing at computer

Stay in the Know

Subscribe now to receive our newsletters.