DHS unveils new AI roadmap

The Department of Homeland Security is launching a slew of new efforts to make use of artificial intelligence for federal operations, with the release of the agency’s inaugural AI roadmap on Monday, designed to add automated efficiency while protecting privacy rights.

The roadmap will guide the agency’s AI use through three lines of effort called for in the Biden administration's AI executive order: responsibly leveraging AI to protect civil liberties; promoting national AI safety and security in the context of critical infrastructure; and leading in AI solution acquisition and development through public and private partnerships. 

DHS Secretary Alejandro Mayorkas and agency CIO Eric Hysen called the roadmap the “most detailed AI plan put forward by a federal agency to date” in a letter prefacing the document.

“What we learn from the pilot projects will be beneficial in shaping how the Department can effectively and responsibly use AI across the homeland security enterprise moving forward,” Mayorkas added in a press release, noting that the “roadmap and pilots will guide our efforts this year to strengthen our national security, improve our operations, and provide more efficient services to the American people, while upholding our commitment to protect civil rights, civil liberties, and privacy.” 

Cyber and physical security remain chief priorities for the agency, with the roadmap calling such efforts "foundational" to safe AI and highlighting the newfound dangers to critical infrastructure operations posed by AI-enabled hacks.

“DHS and the Cybersecurity and Infrastructure Security Agency in particular will continue to work to improve the nation’s overall cyber resilience and to identify and manage risks associated with the misuse of AI/ML technologies,” the document reads.

CISA will assess the use of AI software to help detect cybersecurity vulnerabilities across the agency’s digital networks as well as in other federal civilian government systems, authoring a report for the agency secretary with a list of recommendations based on this initial assessment. 

Part of DHS efforts to address critical infrastructure risk will hinge on greater interagency collaboration and partnership with other sector risk management agencies, namely the National Institute of Standards and Technology, the Department of Justice and the Department of Defense.

Extended partnerships focused on mitigating AI-specific threats will include more than just federal offices. 

Earlier this year, DHS announced a hiring push designed to attract more than 50 new AI specialists to constitute an agencywide AI Corps.

3 new use cases

As part of the roadmap, the agency announced three AI pilot programs that will help to inform department wide policies.

U.S. Citizenship and Immigration Services will apply large language models to train Refugee, Asylum, and International Operations officers on immigration applicants interviews, with the goal of creating personalized training materials that address officers’ specific needs.

A second pilot program at the Federal Emergency Management Agency aims to use generative AI to assist in state, local, territorial and tribal governments’ hazard mitigation planning efforts, with the goal of enabling more communities to submit grant applications to FEMA’s Hazard Mitigation Assistance program. 

The third pilot announced in the roadmap will utilize AI software and large language models at Homeland Security Investigations to enable officers to search for and retrieve relevant case documents, providing them with summarized responses to queries about such information. That pilot’s objective will be to enhance investigative processes in detecting the presence of fentanyl, further combating child sexual exploitation and addressing other criminal enterprises.

“AI, applied strategically, serves as a transformative tool by boosting productivity and improving customer experience,” the roadmap reads.