Talent and training are some of the biggest challenges in artificial intelligence strategy. But what can DOD workers learn from if they embed with private companies?
Talent and training are some of the biggest challenges when it comes to employing artificial intelligence, especially in the Defense Department. But could doing more employee swaps with commercial companies change that?
Brian Drake, AI director for the Defense Intelligence Agency's Science and Technology Directorate said personnel needs could be better satisfied through talent swap or exchange programs that would "embed government employees in private sector entities" but the issue lies with government workers skill sets.
"We have these programs already, called public-private swaps," he said during a GovExec virtual event on May 27. The challenge, however, is "often who will accept the government employee."
Drake said the government is transparent with the fact that they're in the nascent stages of AI development and use, which means the employee who would be swapped out would have eagerness, motivation and a team-player mindset but not the technical skills.
"We're very upfront about [being] in the infancy stages of this. So the employee I'm going to give you is going to be highly motivated, wants to make a lot of change, is going to be very loyal to your business objectives and they're going to be eager to learn," Drake said.
"Will they be able to help you with your Python scripts? No, they probably will not. Will they be able to help you with building a robust data model? Maybe, it depends on who we have."
But Drake said increasing the use of such partnerships could be powerful to developing DOD's AI talent pool.
Fred Allman, NVIDIA's senior director for its U.S. public sector business, made similar points during the panel discussion, but said embracing data science and building training programs for others will be crucial -- and it will have to start with leadership.
"Everybody needs to be ready," Allman said. "To go even faster, while reskilling and training efforts are underway, we must work with public and private partnerships," across academia, startups, government contractors, commercial industry, "to raise the level of awareness and capability of our workforce."
Meanwhile, the Pentagon is working on building out a definitive AI education and training strategy, but worries about being overly prescriptive.
Lt. Gen. Jack Shanahan, the Defense Department's AI chief said the joint effort under the Joint Artificial Intelligence Center is weighing multiple options for better identifying talent, including paying coders the equivalent of language proficiency pay, to organically grow DOD's emerging tech initiatives.
"We want to think hard about doing this right in emerging tech, disruptive tech," he said, naming that getting people who are adept in quantum and blockchain the next five years.
"As part of this early work…[we're] looking at archetypes, what should those archetypes be, and what can we do early in the process to start making some proposals about the training aspect of it and the education piece."
The AI chief said, Shanahan, who leads the Joint Artificial Intelligence Center, during an AFCEA DC virtual event May 21, he sees the AI career fields taking a similar trajectory to cyber and that such details would be developed in the education and training strategy.
For example, the JAIC is figuring out when and how to best identify talent, such as during recruiting -- and making sure they stay in a technology field rather than getting reassigned to an unrelated field -- which often happens to enlisted service members.
Shanahan said future career fields for military service members in AI or machine learning were possible but cautioned about them being too narrow.
"Maybe a little bit of a caution of should it be just AI or should it be something a little bit broader called emerging tech and then you have subsets that would be something like AI, skilled identifiers," he said.
NEXT STORY: NSA warns Russian hackers exploited email flaw