Homeland Security launching investigative AI programs

Secretary of Homeland Security Alejandro Mayorkas announced pilot programs on Monday to integrate artificial intelligence into Homeland Security operations. He hopes the test programs improve national security plans while also protecting civil rights and privacy. “The unprecedented speed and potential of AI’s development and adoption presents both enormous opportunities …and risks we must mitigate,” Mayorkas said. The department points to President Joe Biden’s 2023 executive order on the safe development of AI as the impetus for the projects. The programs will cost $5 million, The New York Times reported.

What will the programs do?

• Homeland Security Investigations plans to use open-source AI “to enhance the efficiency and accuracy” of investigator case summaries, allowing for quicker summary searches and finding reports with relevant contextual information. HSI hopes the pilot will aid in exposing “fentanyl-related networks,” as well as perpetrators and victims of child exploitation crimes, by identifying trend patterns.

• The Federal Emergency Management Agency’s generative AI pilot aims to streamline hazard mitigation planning for “State, Local, Tribal, and Territorial governments.” It hopes the program will aid communities with fewer resources “to become more resilient and reduce disaster risks” while providing more access to grants and funding.

• United States Citizenship and Immigration Services will launch an “interactive application” for training of immigration personnel. The GenAI program would create “dynamic, personalized training materials that adapt to officers’ specific needs” to help with information retention and limit re-training personnel.

How is Mayorkas getting along with Congress since the House impeached him? The House Homeland Security Subcommittee on Border Security and Enforcement will hold a hearing Thursday to examine DHS’ “unlawful use of the CBP One app to mass-parole hundreds of thousands of inadmissible aliens into the United States,” according to the subcommittee.

Dig deeper: Read Josh Schumacher’s report in The Sift on an OpenAI report exploring the threats AI poses.