Defense

Pentagon’s CDAO wraps up ninth iteration of GIDE

CDAO's GIDE 9 successfully demonstrated a “completely vendor-agnostic” data integration layer for the first time.

Anastasia Obis

March 26, 2024 6:46 pm

4 min read

At the end of 2023, the Chief Digital and Artificial Intelligence Office delivered a minimum viable capability for Combined Joint All-Domain Command and Control via a series of global experiments known as GIDE. In 2024, the office wants to make the data collected during GIDE exercises available to those developing the next round of AI models.

This year’s GIDE 9, which was wrapped up last week, successfully demonstrated a “completely vendor-agnostic” data integration layer for the first time, making data extensible across various operational systems. This data can now be fed into the DoD’s development pipeline.

“Now it is ready to start piping over into the development pipeline into Alpha-1 and other capabilities so that we can start learning at scale as an enterprise. That’s one of the things we really want to get after this year is to start making that pipeline permanent, persistent and real so we can start training those models,” said Air Force Col. Matthew Strohmeyer, who leads the GIDE series, during the Center for Strategic and International Studies event Tuesday.

Last month, CDAO’s chief Craig Martell said his goal is to have data mesh in place, which will allow information to flow in a secure manner.

At the strategic level, the GIDE series is testing data mesh services that will ultimately allow combatant commands and the Joint Staff to have data they are able to exchange, giving information advantage on the global scale.

“The data mesh services that we are trying to bring to bear allow us to be able to have data in common between the combatant commands so that one command doesn’t have their kind of program of record that they’re working with that has data in a stovepipe. They may have, for example, some logistics data or munitions data that is relative to that force that they have,” said Strohmeyer.

“In the past, that data wasn’t viewable by another combatant command. But now, because we’re trying to truly globally integrate everything we do, a data mesh service allows us to have that piece of data shared by all the combatant commands. And not just shared via email. It’s shared live.”

At the same time, the data mesh services look different at the tactical level when the services conduct joint fire missions. When it comes to the strategic level, for example, it’s primarily enterprise-level data that is available in the cloud. For tactical-level decisions, data has to be extremely resilient as it needs to withstand operating in environments that will be contested.

During GIDE 9, the team was able to test out data mesh services for the joint operating system.

“We had a true data mesh deployed where all of the data that was used for those warfighting fires decisions existed on every node and the nodes were intelligently routing the data across this kind of mesh network so that if a piece of that mesh went down, it didn’t matter that the data would resiliently repopulate across the mesh and be able to get that information wherever it needed to go and at whatever time,” said Strohmeyer.

Strohmeyer said that while his office is working on integrating the data collected at the strategic and tactical levels into the development phase, which will allow the development of the next round of AI models, they have a “long way to go.”

On the experimentation side, multiple combatant commands conducted a blind test of an AI capability for logistics-related tasks. An example scenario is analyzing the logistics of moving sustainment capability from one location to another. Some of the participants had access to generative AI tools to quickly come up with a recommended path, while some of the participants were coming up with recommendations without generative AI tools.

“The difference was that this wasn’t research organizations that were actual warfighters that were doing this and seeing what worked and what didn’t work as they went through the process,” said Strohmeyer.

This test is one example of how the GIDE series provides Task Force Lima, the CDAO’s initiative to integrate generative AI tools across the DoD, a venue to experiment with large language models.