The U.S. military wants to better understand how its acquisition and contracting workforce can use generative AI to improve performance and is launching a pilot next month to explore these questions.
Jennifer Swanson, deputy assistant secretary of the Army for data, engineering and software, said the effort will highlight how service acquisition and logistics enterprises can leverage generative AI tools to improve contract writing and data analysis. As processes can be made more efficient.
“The pilot is not just about increasing our productivity, which would be great, but also about what other things we can do and what other industry tools are out there that we can use. can leverage or augment that,” Swanson said June 18 at the Defense Ones Tech Summit in Arlington, Va.
The Army is the latest Department of Defense agency to announce efforts to experiment with generative AI. The Air Force and Space Force unveiled their experimental tool — the Unclassified Internet Protocol Generative Pre-Training Transformer, or NIPRGPT — last week. And in 2023, the Navy launched Amelia, a conversational AI program that sailors can use to solve problems or provide tech support.
Swanson said he's optimistic about creative AI's potential, especially for labor-intensive specialties like contract writing and policy where automation could relieve some of the pressure on the Army's workforce.
“In terms of contracts and policy, I think there's a huge return on investment for us,” she said. “Probably [AI] Be able to write a contract one day? We hope. But we have to pilot and test it and make sure everyone is comfortable with it first.
Swanson said the large language model the service will use for the effort is different from systems like ChatGPT, because it's trained on Army data. It will also provide references that indicate where the data it provides originated, a feature that will help the Service verify the authenticity of that information.
The pilot is part of a broader effort within the Army to identify both the pitfalls and opportunities that come with the widespread adoption of AI tools. In March, the service announced a 100-day plan focused on reducing the risk associated with integrating AI algorithms.
As part of that exercise, Swanson said, the military reviewed its spending on AI research and found that testing and security were the two biggest gaps toward fielding these tools more widely. The Service has also identified 32 threats and 66 mitigations it can implement to reduce their impact. Further, it created a creative AI policy that would be applied to the pilot to set the parameters for the effort. This policy includes a requirement to have a “human in the loop”.
Creative AI will lead to the next phase of the pilot effort – a 16-month focus on how to put the technology into practice. The results of this work will inform the Army's fiscal year 2026 budget.
“So the 100-day plan is setting up the situation — where we are — and then the 500-day plan is really about driving it,” he said.
Florent Gruberg, vice president of strategy and optimization at private equity firm AE Industrial Partners, said that as the military goes through these review processes and experiments with AI, it must be transparent with industry about what What she wants and then must move quickly to take advantage of the tools. Companies are developing.
“To me, it's really understanding the framework of what you want to do,” he said during the same panel with Swanson. “Put some boundaries in there and then do it.”
Courtney Albon is C4ISRNET's space and emerging technology reporter. He has covered the US military since 2012, focusing on the Air Force and Space Force. He has reported on some of the Defense Department's most important acquisition, budget and policy challenges.