Defense Department Plans Dedicated AI Development Center Within Six Months

The Defense Department plans within six months to establish a laboratory dedicated to artificial intelligence, which Deputy Defense Secretary Pat Shanahan said is the key to saving billions of dollars by streamlining the Pentagon’s business and computing systems.

Currently under the supervision of Undersecretary of Defense for Research and Engineering Michael Griffin, the Pentagon’s pursuit of AI will eventually fall under the direction of Dana Deasy, a former JP Morgan Chase IT official who should become the Defense Department’s chief information officer (CIO) sometime in May.

iStock Cloud Computing

Shanahan said applying AI to back-office and logistics processes could yield billions of dollars in efficiencies before the controversial technology ever reached the battlefield.

Griffin told lawmakers last week during testimony of plans to establish an AI center of excellence that will consolidate the more than 500 ongoing AI development efforts under one roof. The center will at first seek to streamline the Defense Department logistics, information technology and acquisition systems, which Shanahan said were low-hanging fruit where AI would have an immediate impact.

“It probably won’t surprise you that we don’t always order the right things,” Shanahan said April 24 at a Defense Writers Group breakfast in Washington, D.C. “So, think about the Defense Logistics Agency. Think about all the things we have to manage … supporting all the bases. That is a fantastic place to apply artificial intelligence.”

“As we stand up our center of excellence, we are not going to go out and solve the most vexing problems first,” he added.

AI is one of six major areas in which Shanahan and Griffin believe the United States military has fallen behind near-peer competitors and are determined to flip that script through investment in research. Others include directed energy and hypersonics. 

Instead of powering Terminator robots against enemy forces, the initial applications of artificial intelligence will help speed up and simplify mundane tasks performed by the civilian-dominated agencies that make the Defense Department tick. Shanahan said applying AI to organizations like the Defense Information Systems Agency could satisfy an effort led by House Armed Services Committee Chairman Rep. Mac Thornberry (R-Texas).

Thornberry's proposed “Comprehensive Pentagon Bureaucracy Reform and Reduction Act” would eliminate DISA by Jan. 1, 2021 and transfer its functions to U.S. Cyber Command and other agencies. 

DISA, which has more than 8,000 military and civilian employees, is charged with providing command-and-control and information-sharing capabilities to joint warfighters, national leaders and coalition partners.

Thornberry’s plan also eliminates the Defense Technical Information Center, the Defense Technology Security Administration, the Test Resource Management Center and Washington Headquarters Services.

The bill also would reduce the number of senior executive service-level CIOs in DoD from 60 to five.

So far efforts to reform those organizations – the collective business operations of the Defense Department that Shanahan calls the “Fourth Estate” – has been the job of John Gibson, the Pentagon’s new chief management officer.

Shanahan said the Fourth Estate can be subdivided into three sectors: intelligence, acquisition, and business operations such as health care, pay systems and facilities management. Efficiencies through automation and consolidation are available in all three areas, he said.

Plans are in the works, for instance to standardize acquisition across the Defense Health Agency’s hundreds of medical clinics. DISA likewise should consolidate its data centers as the Pentagon migrates its information technology systems to the cloud, where AI can be applied to find further cost savings.

“The way I tend to think about it is, ‘How do we restructure ourselves so we can be much more productive and much more responsive?’” he said.

More Stories You Might Like