The Department of Homeland Security on Tuesday released the latest version of its Artificial Intelligence Use Case Inventory, reporting 158 active applications of the technology, a significant increase from 67 published last year.
In an explanatory blog post outlining how the agency is approaching artificial intelligence, Eric Heisen, the agency’s CAIO and CIO, outlined 29 AI use cases that have been implemented and 10 that are coming up. It said the AI use cases in question were considered to be at a new level as they had an impact on rights and safety. Oversight established by recent White House guidance. About half of the use cases deployed involved technologies related to facial recognition and facial capture technology, he said.
Use cases for Customs and Border Protection include Babel’s technology, passive body scanners to help detect weapons at border crossings, and autonomous underwater vehicles. Immigration and Customs Enforcement uses facial recognition-based biometric check-in tools, as well as facial recognition technology to identify victims of child exploitation and facial recognition services used in other investigations. I am. The Transportation Security Administration lists AI-enabled Axon body cameras and generative AI systems for employee workflows. Check out our inventory to learn more about these use cases.
The document also mentions an internally generated AI chatbot used by the Cybersecurity and Infrastructure Security Agency.
“Each use case is unique to each practice, including testing performance in real-world situations, maintaining human oversight and accountability, and enabling AI to continuously monitor and mitigate discrimination,” Heisen said in a blog post. “We have confirmed and approved that the requirements are met.” Among other requirements. ” He noted that there is no need to issue a waiver for the required risk management practices. He added that the Office of Management and Budget has approved five short-term compliance deadline extensions for five use cases that impact safety and rights.
DHS says the purpose of this release is to expand its inventory and err on the side of transparency, a decision based on the agency’s governing board’s guidance to reveal as much information as possible. , “even if some details about specific use cases cannot be made public.” I shared it. ” Under the recent National Security Memorandum, AI related to the intelligence community is excluded, but officials plan to release more information on their approach to these use cases in April.
The complete inventory, including deactivated use cases, is available here.
FedScoop continues to report on the accuracy and transparency of AI inventories, including changes to DHS inventories made last year. Researchers at Stanford University also track these inventories, and government watchdogs have previously flagged problems with DHS’s tracking.
