Pentagon’s Joint Common Foundation AI platform is up and running

The Department of Defense launched a new coding platform aimed at helping users across the military build their own artificial intelligence models.

The Joint Artificial Intelligence Center’s Joint Common Foundation (JCF) has reached “initial operating capability” and already has some users in the services, center Director Lt. Gen. Michael Groen said Tuesday, although he did not specify what type of projects or who is involved.

The JCF is meant to be a one-stop-shop for anyone from dabbling data amateurs looking to fill out a slide-deck to full-on machine-learning developers hungry for clean data and an environment to write code. It will play a central role for the developing JAIC, especially as it turns to being an “enabling force” across the DOD rather than working on specific AI projects.

“The JCF is live, we have the tools, we are starting to develop, we are starting to host data, we are starting to host algorithms,” Groen said during the National Defense Industry Association’s inaugural National Security AI Conference and Exhibition. “We hope to grow that into full operating capability.”

The plan is to add a “block upgrade” every month to the platform to expand its data hosting, coding and other capabilities.

“Every month we want to add more services,” Groen said.

Other DevSecOps platforms, somewhat similar to the JCF, exist across the military services, including the Air Force’s Platform One. But Groen said the JAIC’s market for the JCF is made up of those who do not already have access to such a service-designed platform.

User feedback will play a major role in the early development of the platform. The JAIC is using user surveys to solicit initial users and those who would be using the JCF to hear what they will want. Working more closely with the individual services and the many AI offices across the department is a new focus of the JAIC. While initially the center was stood up as an AI fielding office to deliver products in key mission areas, now in its second iteration as the “JAIC 2.0,” it is focused on enabling others to build their own tools.

“We think that is a key tool to broad enablement across the department in the transformation of AI,” Groen said.

The hope is to eventually stitch together a common “data fabric” for enhanced interoperability and usage across the department, Groen added.

The JAIC inked a $106 million deal in August with Deloitte to help build the JCF platform.