Scarlet Dragon exercises, Maven Smart System paving the way for AI adoption, development

Too often, conversations about how the U.S. military can best take advantage of the commercial sector to foster innovation revolve around what’s not working. But the story of how the Army’s 18th Airborne Corps utilized a series of exercises called Scarlet Dragon to develop AI-powered Maven Smart System that is now delivering real value to frontline soldiers, if adopted more widely, could transform future innovation efforts and integration of technologies like AI and software within the DoD.

While many of the lessons learned during the process once again highlighted the importance of flexible approaches to contracting and funding or willingness on the senior leaders’ part to advocate for a program, MSS and Scarlet Dragon showed that “trilingual” leaders — those who can navigate the acquisition and contracting systems as well as technical and operational domains — were critical to the successful integration of MMS.

“What I noticed when we were going through and really coming to understand the story of how MSS and Scarlet Dragon were developed was that the key players all seem to have not just an understanding of what software and AI could do for them, and not just an understanding of how to do artillery fires or how operations really happen.  They understood how to make the acquisition and contracting system work. They understood the boundaries and where they could maneuver,” Emelia Probasco, a senior fellow at Georgetown’s Center for Security and Emerging Technology, told Federal News Network.

“I think sometimes they maneuvered outside of the norm but still within the bounds. And that common thread turns out to be one of our major recommendations, that even though we think it’s very important that all operators are bilingual in the sense that they understand both technology and where technology is going and operations, there should be an ability for a subset of officers to also get that acquisition knowledge and understanding,” Probasco said.

In her latest report, Probasco argues that while not everybody needs to be a trilingual leader, but the DoD leadership should consider adjusting promotion policies and personnel rules to recognize and support these trilingual leaders.

“Having a subset who can work the three in tandem was one of our major recommendations because it seemed to be so consequential to what was achieved in this particular story,” Probasco said.

In June, the Pentagon announced it was rolling out MSS to thousands of users worldwide as part of a major contract with a maximum value of $480 million. The system pulls data from a wide range of sources — existing intelligence databases, satellite intelligence, publicly available data, such as social media feeds or open-source intelligence — and provides a single interface that allows users to access, analyze and make decisions based on the data. The system is also a successful example of the Pentagon’s Combined Joint All-Domain Command and Control concept being put into practice.

Lowering barriers

To onboard new technology companies for MSS development and Scarlet Dragon, the Army used the Secure Unclassified Network (SUNet), which ESC developed several years ago. The network enabled the creation of virtual enclaves — secure, digital environments where groups can collaborate. Solutions like SUNet reinforced the importance of the DoD continuing to lower barriers to entry for new vendors.

“The onboarding was an issue, as I understand it, early in the project. Just getting permission and being able to work on the DoD Information Network is, I’m told, not pleasant,” Probasco said.

“The idea is that as much as we can lower barriers to at least the experimentation, if not operationalization. Let’s just lower the barriers to experimentation so you can get these commercial firms on board trying out their solutions, and then we can worry about some of the other bigger hurdles to cross later. But we can’t even see the art of the possible until we get the barriers low enough to allow these tech companies to help.”

Probasco also said the military needs a repeatable approach for developing and deploying software and algorithms that can be used across different units. “Solutions developed for one unit should be accessible to others if the use cases are fundamentally the same,” she wrote.

The Chief Digital and Artificial Intelligence Office recently announced its Open Data and Applications Government-owned Interoperable Repositories (Open DAGIR) multi-vendor ecosystem, which will provide this standardized repository where various applications and tools can be stored and accessed by the military.

Ultimately, MSS is just one successful example of operationalizing software and AI across the DoD, but the exercise “achieved an efficiency comparable to the best the U.S. military has achieved in recent history — with two thousand fewer staff — should serve as a strong incentive for future responsible experimentation and development.”

Copyright © 2024 Federal News Network. All rights reserved. This website is not intended for users located within the European Economic Area.