Starting in July, the U.S. Army plans to pilot a generative artificial intelligence (AI) program in its acquisition, logistics, and technology division to streamline tasks such as contract writing and data analysis.
Leveraging AI for Army-Related Initiatives
Jennifer Swanson, deputy assistant secretary of the Army for data, engineering, and software, expressed that the initiative will provide insights into how the Army's acquisition and logistics operations could leverage generative AI tools. Swanson emphasized during Defense One's Tech Summit in Arlington, Va., on June 18 that the pilot aims to boost productivity and explore additional industry tools that could be leveraged or integrated into Army operations.
The Army is the latest Defense Department agency to announce an initiative to experiment with generative AI. The Air Force and Space Force previously introduced their experimental tool, the Non-classified Internet Protocol Generative Pre-Training Transformer (NIPRGPT). In 2023, the Navy launched a conversational AI program named Amelia to assist sailors with troubleshooting and provide tech support.
The Army's Large Language Model (LLM) Comes with Citations
Swanson expressed optimism about generative AI's potential, particularly in fields like contract writing and policy, where automation could alleviate burdens on the Army's workforce, highlighting significant potential returns in contracts and policy.
The large language model that the Army will use differs from systems like ChatGPT because it is trained on Army-specific data that will include citations to show the source of its information, aiding in fact-checking for the service. The pilot program represents a broader initiative within the Army to assess widespread AI tool adoption's challenges and potential benefits.
The Crucial Trial Phase
As part of its assessment, Swanson noted that the Army reviewed its expenditures on AI research, identifying testing and security as the primary areas needing improvement for broader deployment of these tools. The Army also pinpointed 32 risks and devised 66 measures to mitigate their effects. Additionally, the Army formulated a policy for generative AI, which will govern the pilot project by setting specific parameters and mandating the presence of a human in the decision-making process.
READ ALSO : Tech Giants Pledge AI Safety Commitments, Introduce "Kill Switch" for Worst Case Scenarios
The generative AI pilot will transition into the subsequent phase of the initiative, a 16-month period focused on the technology's operational implementation. Insights gathered during this phase will guide the Army's budget planning for fiscal year 2026. Swanson explained that the 100-day plan sets the groundwork, establishing the current state, while the 500-day plan focuses on practical implementation.
Florent Groberg, vice president of strategy and optimization at AE Industrial Partners, stressed the importance of transparency from the Army during these evaluation stages and urged the swift adoption of tools being developed by industry partners. During the same panel with Swanson, Groberg emphasized the importance of defining clear objectives and establishing a framework for achieving them, stressing the need to set boundaries and proceed with implementation.
