The US is testing AI to help with immigration, human trafficking and disaster relief efforts

The US Department of Homeland Security (DHS) is launching three $5m pilot programmes to bring artificial intelligence to its agencies. OpenAI, Anthropic and Meta are participating in the project.

Here's What We Know

Under the pilots, the Federal Emergency Management Agency (FEMA) will use generative AI to optimise disaster response planning. Immigration and Customs Enforcement (ICE) will leverage big language models to speed up database searches when investigating child exploitation, human trafficking, and drug trafficking.

US Citizenship and Immigration Services (USCIS) will use AI-powered chatbots to train officers who interview asylum seekers. According to the New York Times, USCIS agents will be able to conduct mock interviews with the AI bots.

DHS said the use of AI will increase detection of criminal networks related to fentanyl distribution and improve identification of criminals and victims of child exploitation.

The agency is forming a task force of at least 50 professionals to coordinate AI projects. However, hiring qualified personnel is not the only challenge. DHS has previously encountered difficulties with AI, including cases of deception by AI-generated deepfakes.

Some DHS divisions, such as Customs and Border Protection, already use AI systems for border surveillance. But the new pilot programmes will be the first large-scale use of large language models domestically. DHS plans to report on the results of the experiments by the end of 2024.

Source: The Verge