Generative AI Presents Key Opportunities and Challenges for Defense Department
• Article
October 20, 2023—Generative artificial intelligence (AI) has garnered popular attention for content-creation applications such as ChatGPT. For Department of Defense (DoD) agencies, however, the technology presents special opportunities and risks. A paper released by the Software Engineering Institute (SEI) outlines where the DoD might benefit from generative AI, its risks, and research challenges in this burgeoning field.
The paper captures the findings of a June workshop convened by the Under Secretary of Defense for Research and Engineering (USDR&E) and the SEI to elicit DoD use cases for generative AI. Attendees included members of the armed services and federally funded research and development centers.
The workshop found that the DoD stood to benefit from generative AI in the fields of business, wargaming, coding, and simulation. However, the defense context raises the stakes for the common risks of generative AI—AI hallucinations or fabrications, mis- and dis-information campaigns, data poisoning, and prompt engineering attacks. Classification concerns for LLM-constructed documents and AI-ready computing resources in tactical environments are particular concerns for the DoD.
“Generative AI is quickly changing the way individuals and organizations work, and with those changes come potential impacts to operations of the DoD,” said Matt Gaston, director of the SEI’s AI Division, which hosted the June event. “This timely workshop brought together government labs and military services to level set, early on, both the positive impacts and risks of adopting generative AI.”
Read the paper in the SEI’s Digital Library. Learn more about the SEI’s work in AI engineering, responsible AI, and other aspects of this fast-changing technology on the SEI Blog.
