The Department of Homeland Security (DHS) is rolling out three AI pilot programs at a cost of $5 million across its three agencies. of new york times report. Through a partnership with OpenAI, Anthropic, and Meta, DHS will test AI models to assist agents in a wide range of tasks, including investigating child sexual abuse materials, training immigration officials, and creating disaster relief plans.
As part of the AI pilot, the Federal Emergency Management Agency (FEMA) will use generative AI to streamline hazard mitigation planning processes for local governments. Homeland Security Investigations (HSI), an agency within Immigration and Customs Enforcement (ICE) that investigates child exploitation, human trafficking, and drug smuggling, uses large-scale language models to quickly store large amounts of data. Search and summarize research reports. Additionally, U.S. Citizenship and Immigration Services (USCIS), which conducts introductory screening for asylum seekers, plans to use chatbots to train its staff.
D.H.S. announcement I don't know much about the details, but times The report provides several examples of what these pilots might look like in practice.by times, USCIS asylum agents use chatbots to conduct mock interviews with asylum seekers. Meanwhile, HSI investigators will be able to more quickly search suspect details in internal databases, which DHS says will „lead to increased detection of fentanyl-related networks“ and „identify perpetrators and victims of child exploitation crimes.“ It may be useful.'' ”
To accomplish this, DHS is building an „AI force“ of at least 50 people. In February, DHS Secretary Alejandro Mayorkas traveled to mountain viewCalifornia, famous for Google's headquarters, has begun hiring AI talent, appealing to potential candidates by emphasizing that the department is „incredibly“ open to remote workers.
Hiring enough AI experts isn't the only hurdle for DHS.as times He said DHS's use of AI has not always been successful, and there have been times when agents have been tricked into investigations by AI-generated deepfakes. February report The Government Accountability Office investigated two AI use cases within the department and found that DHS did not use reliable data in one investigation. In another case, DHS claimed it did not rely on AI at all. Beyond DHS, there are numerous documented cases where ChatGPT has spewed false results. These include instances where lawyers submitted briefs citing non-existent examples of AI models. completely finished.
However, this expansion is not DHS' first foray into AI. Part of a surveillance tower used by Customs and Border Protection (CBP) to monitor the U.S.-Mexico border. Made by Anduril, uses AI systems to detect and track “objects of interest” as they move through the rugged terrain of border areas. CBP wants this fully integrate By 2034, we plan to build a network of AI-powered watchtowers. Authorities also plan to use AI to monitor official border crossings. Last year, CBP awarded a $16 million contract to a high-tech travel company founded by former Commissioner Kevin McAleenan. Build an AI tool to scan for fentanyl At the port of entry.
However, DHS' new AI pilot program relies on large-scale language models rather than image recognition and will be used primarily domestically rather than at the border. DHS plans to report trial results by the end of the year.