DOD faces workforce challenges for AI test and evaluation

The Pentagon needs to test and evaluate new AI-enabled systems to make sure they’re up to snuff before they end up in the hands of troops.
(Getty Images)

The Pentagon needs to test and evaluate new artificial intelligence-enabled systems to make sure they’re up to snuff before they end up in the hands of troops. However, the department will be challenged to find more personnel who are qualified to do this type of work, according to an official in charge of leading these efforts.

Achieving “AI assurance” is a key pillar of the U.S. military’s plans to proliferate these types of technologies across the force, from the back office to the battlefield.

“Workforce is critical for this,” said Jon Elliott, chief of test and evaluation (T&E) at the Pentagon’s new Chief Data and Artificial Intelligence Office (CDAO).

“Historically, test and evaluation experts in the DOD are always in high demand, and the DOD has always had a shortage of them. In addition to that, adding the expertise in AI on top of that — which is also in high demand both in industry and the DOD — it becomes … in my particular area of T&E of AI a very hard role to find someone who already fills [the job requirements] or has that background,” he said Tuesday during a panel at the Pentagon’s Digital and AI Symposium.


More education and training will be needed, he said, whether it’s for people the DOD recruits from outside the government or personnel who are already on the payroll.

“It’s either bringing in people with AI skill sets into the government and helping to train and discuss how the DOD does test — which is different than how industry does test — as well as … bringing in the test experts from the DoD and test ranges and educating them on how AI works and the underlying mathematics behind [things] like neural nets that influence the decisions they make,” Elliott said.

The way the U.S. military tests more traditional systems won’t cut it when it comes to putting AI capabilities through their paces.

DOD weapons testers need to “understand why the current paradigm of like driving a truck, you know, 5,000 miles around a track determines its reliability does not work for an AI system or an autonomous ground vehicle that the Army is purchasing or the DoD is purchasing,” he said.

Jon Harper

Written by Jon Harper

Jon Harper is Managing Editor of DefenseScoop, the Scoop News Group’s newest online publication focused on the Pentagon and its pursuit of new capabilities. He leads an award-winning team of journalists in providing breaking news and in-depth analysis on military technology and the ways in which it is shaping how the Defense Department operates and modernizes. You can also follow him on Twitter @Jon_Harper_

Latest Podcasts