Leonel Garciga, chief information officer of the U.S. Army and a 2024 Wash100 awardee, has issued a memorandum detailing new policy guidelines for the development, deployment and use of generative artificial intelligence and large language models across the service branch.
Published on Thursday, the document instructs generative AI developers and system owners to comply with existing legal and security regulations, obtain necessary approvals for processing sensitive and classified data and ensure that users can effectively determine which systems rely on generative AI software.
Developers must also instill processes to document training data sources, test software controlled environments and make sure generative AI applications are authorized to operate within the Department of Defense Information Network.
Army commands are encouraged to develop appropriate processes to evaluate both the risks and benefits of generative AI tools rather than banning the use of these applications.
Command developers, system owners and users must also apply appropriate risk assessment frameworks for generative AI software.