• To reap the benefits of AI, the Defense Department must first tackle challenges with people, processes and infrastructure. Credit: Laurent T/Shutterstock
     To reap the benefits of AI, the Defense Department must first tackle challenges with people, processes and infrastructure. Credit: Laurent T/Shutterstock

Three Steps to Getting AI Right at the Pentagon

July 24, 2019
By Steve Orrin

An effective AI framework demands new thinking and processes.

When it comes to artificial intelligence (AI), the Department of Defense (DOD) has put a firm stake in the ground. The department’s AI strategy clearly calls for the DOD “to accelerate the adoption of AI and the creation of a force fit for our time.”

Yet as the department looks to embrace emerging AI applications, such as video analysis and change detection, sensor fusion and predictive analytics, it must first address the people, processes and technology needed to reap the benefits of these applications. Developing an effective framework for what some consider to be “the single most influential human innovation in history” requires transforming traditional thinking and processes.

That’s not necessarily easy to do, especially in a culture that tends to nourish a climate of restricted, need-to-know information sharing. Here are three steps the organization can take to meet these challenges and build a solid foundation for AI success.

Gather the right people

An effective AI framework requires a workforce that has experience with the tools and technologies as well as expertise with the appropriate algorithms or data frameworks. These people already exist within the DOD; the challenge is bringing them together on the same task force.

The right people can help break down organizational silos. From an institutional perspective, this may be the largest challenge of them all. Data drives AI, and while these silos were historically put into place to secure sensitive data, AI-powered cybersecurity systems can ensure effective security standards and help create better prevention and recovery strategies.

Often we forget that while AI and machine learning answer hard questions, it is really more important to answer the right questions. The ability to leverage AI to identify and differentiate between entities, objects or persons is central to its users. The technology needs to be able to determine if the thing in front of it is harmless or a threat, a friend or foe. This is not a one-and-done process. The models require a constant feedback loop. Everyone from troops in the field to managers at home must be able to understand how to feed information into the system so that system becomes smarter over time.

Look to the commercial sector

The DOD recognizes that the commercial space can offer the department a path toward AI. Indeed, the AI strategy notes that the department “cannot succeed alone” and calls for close collaboration with “non-traditional centers of innovation in the commercial sector.”

The DOD should consider adopting an 80/20 model—look to take 80 percent of an AI solution from a commercial environment, then do the 20 percent modification to support the department’s specific needs. This approach may be challenging but being open to collaboration with the commercial sector would greatly accelerate AI adoption.

Create a framework that supports AI

Of course, building a technology platform upon which AI can thrive cannot be done using legacy technologies. The Pentagon needs to invest in advanced, flexible and agile solutions to be successful. These include software-defined infrastructure, edge-to-cloud architectures and data analytics technologies. It’s important to have the ability to create the types of networks that support collecting and processing data from the edge.

Sensor fusion is another key technology that demands edge-computing capabilities. Sensor fusion may involve multiple cameras, audio sensors, electromagnetic signatures, radar and more. All of these have to be collected and analyzed in real time to gain a complete situational awareness of a battlefield or training scenario.

These capabilities require an architecture that can link together Internet of Things-style devices, creating peer-to-peer and mesh networks with gateways and aggregators. They demand an edge-to-edge and edge-to-cloud data-centric approach to IT modernization.

The Time is Now

In the perilous situations our military personnel can find themselves, being able to process information at the point of contact in the field for immediate, real-time insights can be critical to the success of their missions. AI has the power to provide them with the knowledge they need to make better decisions more quickly than ever before.  

The technology is no longer the stuff of science fiction. It exists today and is readily available. The DOD just needs to ensure it has the right people, partners and frameworks to make it a reality.

Steve Orrin is the chief technology officer, Intel Federal.


Enjoyed this article? SUBSCRIBE NOW to keep the content flowing.


Share Your Thoughts: