Pentagon's AI chief: lack of enterprise cloud 'has slowed us down'

Not having an enterprisewide cloud computing platform has "slowed down" the Pentagon's artificial intelligence efforts, said Lt. Gen. Jack Shanahan, who leads the Defense Department's Joint AI Center.

"The lack of an enterprise solution has slowed us down," but "the gears are in motion," Shanahan said during an AFCEA DC virtual event May 21 on artificial intelligence and machine learning in DOD.

Shanahan said the JAIC is using an "alternate platform" to host a newer anti-COVID effort, Project Salus, a data aggregation platform that uses predictive modeling for equipment needed by front-line workers. That alternate platform was used for the controversial Project Maven, an effort to use AI to tag and interpret images that some – including many in the Silicon Valley workforce – felt was an initial step toward developing lethal AI applications.

Shanahan said the project should be on the Joint Common Foundation, an enterprise-wide and multi-cloud environment. The JAIC was slated to be an early adopter of the delayed Joint Enterprise Defense Infrastructure program – the $10 billion DOD-wide cloud acquisition that was won by Microsoft last October and has been mired in legal battles ever since.

"It's set us back, there's no question about it but we now have a good plan to account for the fact that it will be delayed potentially many more months," he said.

That plan includes using an interim solution at Hanscom Air Force Base, pivoting to the Air Force's Cloud One environment, and using Microsoft Azure and Amazon Web Services capabilities.

"We're kind of covering our bases with [Microsoft] Azure, AWS -- I will never get into a company discussion, I'm agnostic," he said. "I just need an enterprise cloud solution. If we want elastic compute [capability], if we want to make worldwide updates to all these algorithms in the space of minutes not in the space of months running around gold discs, we've got to have an enterprise cloud solution."

The JAIC just forged a five-year $800 million task order for joint warfighter operations to assist in decision-making and data analysis. That deal with Booz Allen was executed via the Alliant 2 contract vehicle at the General Services Administration.

"That is our big focus area for the next year," Shanahan said. "This is showing that we are serious about taking AI and putting it into warfighting operations" that include assisting in emergency responses by converting text and video data in actionable information, and accelerating data streams from sensors to shooters.