As the Department of the Navy looks to enable better and faster decisions for warfighters, Chief Information Officer (CIO) Jane Rathbun said the service branch is infusing AI at the edge to improve decision-making.
At the CyberScape Summit hosted by GovCIO Media & Research on March 7, Rathbun explained how AI capabilities are “critical” for warfighters to achieve mission objectives.
“AI, machine learning, natural language processing, all of those capabilities are critical to managing the massive amounts of data telemetry that come off of our systems that are about cybersecurity,” Rathbun said, adding, “Critical to the cybersecurity and critical in general.”
The CIO said that these tools help to sift through the data, and then make it “more agile and effective” to help warfighters know where to look for threats.
“We’re not going to move all of our data to the edge,” Rathbun said. “We should be, and we are, talking about training models where I’ve got big data stores that I know are secure because they’re in the mothership – more secure than they would potentially be at the edge – and then just moving that training model to the edge, allowing it to consume new data from sensors and giving warfighters solutions, decisions, options, more quickly than we have before. I do think that AI, machine learning, all of these analytic capabilities will be key.”
However, in order to best make use of those analytic tools, Rathbun said that the Department of Defense (DoD) must first ensure that the data is properly labeled, managed, and usable.
The DoD will need to mingle classified and unclassified data to train AI models, Rathbun said, emphasizing that “data maneuverability” is crucial.
“AI is the new cloud,” she said. “Everyone’s talking AI, but where I go back to is, ‘Well, how do we enable an environment that can optimize the utilization of AI? … How can we go back and rethink how we store data, when we classify our data, and how we manage it, so that we can take advantage of these things like large language modeling?’”
Rathbun said that idea can be “frightening” for people because the DoD’s entire ecosystem relies on the idea of different networks based on the classification of data. However, she said the department has to “change that paradigm if we’re really going to leverage these tools and have more prepared data at the edge versus just staying at the edge.”
“I can push all this data and have the warfighter at the edge integrate it and make sense of it and then prepare options for his commander to make decisions,” Rathbun said, adding, “Or I can understand the kinds of decisions that need to be made and start training our models to produce data that will help them make those decisions … more quickly.”
