Everyone talks about artificial intelligence, but identifying any actual voluntary military applications can be difficult.
“If you have data issues or have problems with data issues, it's probably best suited to AI,” said Angela Sheffield, an internationally recognized expert in the nuclear non-proliferation and applications of AI for national security.
Sheffield is cited to “revolutionize” the National Nuclear Non-Proliferation Agency's National Nuclear Non-Proliferation Agency through innovative AI research and development. So, as a senior program manager in AI and Data Science, Sheffield developed a next-generation tool to detect early indicators of illegal nuclear weapon development.
Currently, Sheffield is taking on a new role as director of the AI program at SAIC, the leading system integrator and solution provider for federal and defense applications. The former Air Force intelligence officer sees many opportunities to guide AI to many defense requirements, and says starting is often the most difficult part.
“We have a legacy system forever,” Sheffield says. “We always have fragmented, siloed data repositories. These are not what we can hope for.”
But they don't have to be a barrier to automation either. Whether you are tackling complex issues such as all domain commands and control of combined joints, key initiatives to modernize weapon systems, or efforts to automate tasks, collection, processing, exploitation, and dissemination (TCPED) (TCPED), mundane tasks with routine tasks can be automated to reduce the cognitive load of humans.
“There are many other applications ripe for modernization and innovation, AI opportunities,” Sheffield said. And it doesn't have to be a tough, super complicated use case: business operations and other everyday operational tasks “It's a really great opportunity for us to leverage AI. Automation can free our airlines and other service members and civilians… it can tackle the challenges faced by DOD today.”
In some respects, it's starting to happen. It gains footing with AI-driven capabilities integrated into Enterprise email and other collaboration activities.
“We're beginning to expect that as part of the service we get from businesses,” Sheffield said. AI also added that it can support efficiency in business operations and mission execution in “meeting computing requirements and managing a variety of data sources.”
This is where integrators are particularly valuable. “SAIC is part of bringing these solutions to the Air Force and the rest of the joint forces, with concepts like stove pipes or data layers that interconnect fragmented data systems,” she said.
Once data can be shared across systems, “everyone can get a single site image or a single understanding of all the resources captured in these different repositories,” she says, enabling automated processes supported by AI, enhanced analysis, and informed accelerated decision-making.
Users aren't always easy to purchase automation, she said. You need to gain trust and build to ensure that users gain confidence in their intelligent systems. They need to make sure the software works, Sheffield said, and “to understand whether there is performance in the intended envelope and how it works.” And they need to be certain that the AI is not generating false “hatography,” she said.
The AI must be the main driver for enabling CJADC2. Because without it, the dataset would be too large, and the problem is too harsh, so maintaining the advantages of information at the speed of modern warfare. CJADC2 requires real-time sharing of data across services, national and digital boundaries.
It means overcoming legacy IT load blocks and information systems that cannot speak to each other. Interoperable databases and AI-driven automation are part of the solution. “CJADC2 occurs as a result of its modernization in a more powerful way than what we began to see in our pilot demonstrations,” predicts Sheffield.
For example, Indo-Pacific Command's Joint Fire Network provides real-time, viable threat data for joint, partners, and alliances with battle management systems. SAIC is involved in its pilots, and Sheffield foresees a more AI-driven implementation, such as “closing the kill chain faster and achieving those successes.”
Different systems and technologies, often built with unique technologies, need to be integrated to make them work. “That's where integrators like SAIC can help,” she said.
As a federal program manager, she recalled: “I often relied on contractors and performers to provide that visibility. Commercial partners “have helped me with visibility into what was going on between ministries.”
That's exactly the value Sheffield brings to his work at SAIC. “Looking at our multi-mission portfolio, bringing the best solution for DoD missions is something they can rely on us,” she said.
