The Pentagon has never undergone a clean financial audit. In fact, prior to 2018, the Department of Defense had never completed a departmentwide audit — and in that first exercise, the military’s leadership was not surprised by the results.
However, the head of the Pentagon’s new artificial intelligence capability believes AI could play a key role in helping the DOD shore up its accounting and make a clean audit more than just a pipe dream.
While DOD’s Joint AI Center (JAIC) is chartered primarily to establish and coordinate departmentwide mission-critical AI initiatives — like its ongoing efforts in predictive maintenance, and humanitarian assistance and disaster relief — program lead Lt. Gen. Jack Shanahan said the department has plenty of manual processes, like financial accounting, that can benefit from robotic process automation (RPA) to drive greater back-office efficiencies by augmenting manual, human work with algorithms and bots.
“It’s not headline-grabbing in terms of big AI projects, but it may be where some of the most efficiencies can be found. … This is where early gains are being realized in AI,” Shanahan testified Tuesday before a Senate Armed Services subcommittee. He added that RPA applications can have a “much shorter-term return on investment” than DOD’s other high-profile AI applications.
JAIC has had very “early discussions” with the Pentagon’s chief management officer and chief data officer to forecast what the department’s widespread adoption of RPA might look like.
Until a couple of months ago, Shanahan said, “we weren’t even concentrating on this smart automation. I’m now trying to figure out how I stand up a small office just focused on that.” JAIC probably won’t lead that effort, he said, but it will probably play a role in “leading others to find out how to incorporate there into their back-office functions.”
Shanahan pointed to finance, at the center of the Pentagon’s pursuit of a clean audit, as the first back-office function DOD should look to augment with AI, helping personnel “do much more work than they were able to do with [existing] tools, which in some cases are far too old, manual, laborious.”
When it comes to a bot against a human, he said, “there’s no question who gets to the finish line faster.” But humans shouldn’t be worried about being replaced by robots — “There’s not enough people to go around,” he said.
The main challenge in tackling financial auditing across the largest enterprise in the world will be scoping it, Shanahan said — “finding out how big this is to go after the audit.”
DOD Chief Data Officer Michael Conlin is helping explore this problem from the data side, which is the critical base on which AI stands, Shanahan added. “There are big decisions made in the department with data done in a very manually intensive way. If those decisions can be made faster and better, that, of course, is something that the leadership of the department is interested in.”
Shanahan anticipates the success of back-office RPA will grow as more personnel work with it and grow to understand and embrace. He calls it the “flywheel effect.”
“Once a few people begin to understand what smart automation does, it will begin to catch on,” he said. “But nobody believes it yet because they haven’t had the benefit of seeing it work. … It’s the ‘show me’ piece.”
AI is a major priority for the Pentagon in its fiscal 2020 budget proposal. It asked appropriators for $208 million to scale JAIC, part of a $927 million request to boost the development of AI and machine learning for use in the field. The Pentagon also released the unclassified version of its AI strategy last month.
Some Silicon Valley tech firms have shown reluctance in working with the Pentagon on its AI initiatives because of the ethical implications in being connected to warfighting and lethality. But financial and audit applications seem like a sort of neutral ground in which DOD can attract those innovative commercial companies as partners.