The Department of Homeland Security (DHS) will test the use of artificial intelligence (AI) in three pilot projects, the department announced Monday.
The department will test AI to help investigations focused on detecting fentanyl and combatting child sexual exploitation, along with tests that will be carried out under Federal Emergency Management Agency (FEMA) and U.S. Citizenship and Immigration Services (USCIS).
FEMA will use AI to help communities plan for and develop mitigation plans to build resilience and minimize risks, while USCIS will use AI to “improve immigration officer training.”
“The unprecedented speed and potential of AI’s development and adoption presents both enormous opportunities to advance our mission and risks we must mitigate,” Secretary of Homeland Security Alejandro Mayorkas said in an announcement.
Along with the three pilot projects, DHS announced a roadmap of principles to guide its deployment of AI initiatives.
The roadmap calls for DHS to ensure its use of AI “fully respects” privacy and civil rights and is “rigorously tested to avoid” bias, privacy harms, and other risks. It also urges the agency to promote nationwide AI safety and security, including protecting U.S. cyber networks and critical infrastructure.
The roadmap also calls for DHS to lead in “strong cohesive partnerships,” including with the private sector, local governments, and research institutions.
The AI plan builds on the DHS’s announcement last month that it is hiring 50 AI technology experts.
DHS will spend $5 million on the pilot program announced Monday, The New York Times reported.
The department will use AI models, like OpenAI’s ChatGPT, to assist with the new tests, and is also partnering with Anthropic, Meta, Microsoft, Google and Amazon as it launches the pilot programs, the Times reported.