DHS tests AI for immigration officer training and disaster planning

The Department of Homeland Security (DHS) announced Monday that it will test the use of artificial intelligence (AI) in three pilot projects.

In addition to testing conducted under the Federal Emergency Management Agency (FEMA) and U.S. Citizenship and Immigration Services (USCIS), the department conducts investigations focused on detecting fentanyl and combating child sexual exploitation. We plan to test AI to assist.

FEMA is leveraging AI to help communities plan and develop mitigation plans to increase resilience and minimize risk, and USCIS is leveraging AI to “improve immigration officer training.” Masu.

“The unprecedented speed and potential of AI development and deployment presents both tremendous opportunities to advance our mission and risks that must be mitigated,” Department of Homeland Security Secretary Alejandro Mayorkas said in a statement. It’s bringing that,” he said.

Along with three pilot projects, DHS released a roadmap of principles to guide the deployment of the AI ​​initiative.

The roadmap requires DHS to ensure that its use of AI is “fully respectful” of privacy and civil rights and has been “rigorously tested to avoid bias, privacy violations, and other risks.” We are asking you to confirm. It also calls on the agency to advance AI safety and security nationwide, including protecting America’s cyber networks and critical infrastructure.

The roadmap also calls for DHS to lead “strong collaborative partnerships” with the private sector, local governments, research institutions and others.

The AI ​​plan builds on DHS’ announcement last month to hire 50 AI technology experts.

The New York Times reports that DHS plans to spend $5 million on the pilot program announced Monday.

The department plans to use AI models such as OpenAI’s ChatGPT to support the new tests, and is also partnering with Anthropic, Meta, Microsoft, Google and Amazon to launch the pilot program, the Times reported. .