In a recent discussion at the Surface Navy Association’s 36th National Symposium, Alexandra Landsberg, Division Director of Mathematics, Computers, and Information Science at the Office of Naval Research, shed light on the administrative applications of artificial intelligence (AI) within the Navy. While AI’s capabilities in unmanned systems are well-known, Landsberg emphasized its potential in areas like logistics, asset management, and mission planning.
Landsberg highlighted the opportunity for AI to address operational challenges related to logistics, drawing parallels with industry giants like Amazon and FedEx, which efficiently manage warehouses and optimize scheduling. Leveraging AI approaches from such commercial players could modernize Navy shipyards, enhancing readiness by optimizing resource availability.
Another administrative application discussed was the use of large language models like ChatGPT for searching large documents, producing summaries, and generating accurate forms. However, Landsberg acknowledged the challenge of security and trust when utilizing generative AI within the Defense Department.
To overcome security concerns, she emphasized the need for the Navy to develop secure environments and ensure the security of training data for AI models. Trust and understanding are crucial, requiring collaboration between humans and AI systems. Landsberg stressed the importance of testing, evaluation, verification, and validation, urging experimentation in real-world scenarios to ensure practical applicability.
In conclusion, Landsberg emphasized the necessity of a collaborative ecosystem involving the military, industry, and academia to provide rigorous measures and guarantees for AI. The AI ecosystem’s success, she noted, relies on scaling up from experiments to operational use, requiring collaboration across various sectors and early testing within the Navy fleet.
